243 12 5MB
English Pages 270 Year 2006
Carbon Nanotubes
NATO Science Series A Series presenting the results of scientific meetings supported under the NATO Science Programme. The Series is published by IOS Press, Amsterdam, and Springer in conjunction with the NATO Public Diplomacy Division
SubSeries I. II. III. IV.
Life and Behavioural Sciences Mathematics, Physics and Chemistry Computer and Systems Science Earth and Environmental Sciences
IOS Press Springer IOS Press Springer
The NATO Science Series continues the series of books published formerly as the NATO ASI Series. The NATO Science Programme offers support for collaboration in civil science between scientists of countries of the EuroAtlantic Partnership Council. The types of scientific meeting generally supported are “Advanced Study Institutes” and “Advanced Research Workshops”, and the NATO Science Series collects together the results of these meetings. The meetings are coorganized by scientists from NATO countries and scientists from NATO’s Partner countries – countries of the CIS and Central and Eastern Europe. Advanced Study Institutes are highlevel tutorial courses offering indepth study of latest advances in a field. Advanced Research Workshops are expert meetings aimed at critical assessment of a field, and identification of directions for future action. As a consequence of the restructuring of the NATO Science Programme in 1999, the NATO Science Series was reorganized to the four subseries noted above. Please consult the following web sites for information on previous volumes published in the Series. http://www.nato.int/science http://www.springer.com http://www.iospress.nl
Series II: Mathematics, Physics and Chemistry – Vol. 222
Carbon Nanotubes: From Basic Research to Nanotechnology edited by
Valentin N. Popov Faculty of Physics, University of Sofia, Bulgaria and
Philippe Lambin Département de Physique, Facultés Universitaires NotreDame de la Paix, Namur, Belgium
Proceedings of the NATO Advanced Study Institute on Carbon Nanotubes: From Basic Research to Nanotechnology Sozopol, Bulgaria 2131 May 2005 A C.I.P. Catalogue record for this book is available from the Library of Congress.
ISBN10 ISBN13 ISBN10 ISBN13 ISBN10 ISBN13
1402045735 (PB) 9781402045738 (PB) 1402045727 (HB) 9781402045721 (HB) 1402045743 (ebook) 9781402045745 (ebook)
Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands.
www.springer.com
Printed on acidfree paper
All Rights Reserved © 2006 Springer No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands.
TABLE OF CONTENTS
Preface.……………...………………………………………………………... xi Scientific Committee……..………………………………………………… xiii Part I. Synthesis and structural characterization Arc discharge and laser ablation synthesis of singlewalled carbon nanotubes B. Hornbostel, M. Haluska, J. Cech, U. Dettlaff and S. Roth ………………….1 Scanning tunneling microscopy and spectroscopy of carbon nanotubes L. P. Biró and Ph. Lambin …………………………………………………….19 Structural determination of individual singlewall carbon nanotube by nanoarea electron diffraction E. Thune, D. Preusche, C. Strunk, H. T. Man, A. Morpurgo, F. Pailloux and A. Loiseau…..……………………………………………………………..43 The structural effects on multiwalled carbon nanotubes by thermal annealing under vacuum K. D. Behler, H. Ye, S. Dimovski and Y. Gogotsi ………………………..……45 TEM sample preparation for studying the interface CNTscatalystsubstrate M.F. Fiawoo, A. Loiseau, A.M. Bonnot, A. Iaia, V. Bouchiat and J. Thibault ………………..………………………………………………47 A method to synthesize and tailor carbon nanotubes by electron irradiation in the TEM R. Caudillo, M. JoséYacaman, H. E. Troiani, M. A. L. Marques and A. Rubio ……………………………………………………………..……49 Scanning tunneling microscopy studies of nanotubelike structures on the HOPG surface I. N. Kholmanov, M. Fanetti, L. Gavioli, M. Casella and M. Sancrotti ............51 Influence of catalyst and carbon source on the synthesis of carbon nanotubes in a semicontinuous injection chemical vapor deposition method Z. E. Horváth, A. A. Koós, Z. Vértesy, L. Tapasztó, Z. Osváth, P. Nemes Incze, L. P. Biró, K. Kertész, Z. Sárközi and A. Darabont ………...53 PECVD growth of carbon nanotubes A. Malesevic, A. Vanhulsel and C. Van Haesendonck …………..……………55 v
vi Carbon nanotubes growth and anchorage to carbon fibres Th. Dikonimos Makris, R. Giorgi, N. Lisi, E. Salernitano, M. F. De Riccardis and D. Carbon ...................................................................57 CVD synthesis of carbon nanotubes on different substrates Th. Dikonimos Makris, L. Giorgi, R. Giorgi, N. Lisi, E. Salernitano, M. Alvisi and A. Rizzo ........................................................................................59 Influence of the substrate types and treatments on carbon nanotube growth by chemical vapor deposition with nickel catalyst R. Rizzoli, R. Angelucci, S. Guerri, F. Corticelli, M. Cuffiani and G. Veronese ................................................................................................61 Non catalytic CVD growth of 2Daligned carbon nanotubes N. I. Maksimova, J. Engstler and J. J. Schneider ..............................................63 Pyrolytic synthesis of carbon nanotubes on Ni, Co, Fe/ɆɋɆ41 catalysts K. Katok, S. Brichka, V. Tertykh and G. Prikhod’ko ……………………..…...65 A Grand Canonical Monte Carlo simulation study of carbon structural and adsorption properties of inzeolite templated carbon nanostructures Th. J. Roussel, C. Bichara and R. J. M. Pellenq …………………………...…67 Part II. Vibrational properties and optical spectroscopies Vibrational and related properties of carbon nanotubes V. N. Popov and Ph. Lambin ………………………………………………….69 Raman scattering of carbon nanotubes H. Kuzmany, M. Hulman, R. Pfeiffer and F. Simon …………………….…….89 Raman spectroscopy of isolated singlewalled carbon nanotubes Th. Michel, M. Paillet, Ph. Poncharal, A. Zahab, J.L. Sauvajol, J. C. Meyer and S. Roth ……………………..……………………………….121 Part III. Electronic and optical properties and electrical transport Electronic transport in nanotubes and through junctions of nanotubes Ph. Lambin, F. Triozon and V. Meunier …………………………………….123 Electronic transport in carbon nanotubes at the mesoscopic scale S. Latil, F. Triozon and S. Roche ……………………………………………143 Wave packet dynamical investigation of STM imaging mechanism using an atomic pseudopotential model of a carbon nanotube Géza I. Márk, Levente Tapasztó, László P. Biró and A. Mayer ..……….…...167 Carbon nanotube films for optical absorption E. Kovats, A. Pekker, S. Pekker, F. Borondics and K. Kamaras .....................169
vii Intersubband exciton relaxation dynamics in singlewalled carbon nanotubes C. Gadermaier, C. Manzoni, A. Gambetta, G. Cerullo, G. Lanzani, E. Menna and M. Meneghetti ..........................................................................171 Peculiarities of the optical polarizability of singlewalled zigzag carbon nanotube with capped and tapered ends O. V. Ogloblya and G. M. Kuznetsova ……………….……………………...173 Thirdorder nonlinearity and plasmon properties in carbon nanotubes A. M. Nemilentsau, A. A. Khrutchinskii, G. Ya. Slepyan and S. A. Maksimenko …………………………..…………………………175 Hydrodynamic modeling of fast ion interactions with carbon nanotubes D. J. Mowbray, S. Chung and Z. L. Miškoviü ………….……………………177 Local resistance of singlewalled carbon nanotubes as measured by scanning probe techniques B. Goldsmith and Ph. G. Collins …………………………………………….179 Band structure of carbon nanotubes embedded in a crystal matrix P. N. D'yachkov and D. V. Makaev ……………………….…………………181 Magnetotransport in 2D arrays of singlewall carbon nanotubes V. K. Ksenevich, J. Galibert, L. Forro and V. A. Samuilov ……….…………183 Computer modeling of the differential conductance of symmetry connected armchairzigzag heterojunctions O. V. Ogloblya and G. M. Kuznetsova ……………………….……………...185 Part IV. Molecule adsorption, functionalization and chemical properties Molecular Dynamics simulation of gas adsorption and absorption in nanotubes A. Proykova ……...…………………………………………………………..187 Firstprinciples and molecular dynamics simulations of methane adsorption on graphene E. Daykova, S. Pisov and A. Proykova ……………………….……………...209 Effect of solvent and dispersant on the bundle dissociation of singlewalled carbon nanotubes S. Giordani, S. D. Bergin, A. Drury, É. N. Mhuircheartaigh, J. N. Coleman and W. J. Blau ……………………………………………….211 Carbon nanotubes with vacancies under external mechanical stress and electric field H. Iliev, A. Proykova and F.Y. Li .…………………………………………..213
viii Part V. Mechanical properties of nanotubes and composite materials Mechanical properties of threeterminal nanotube junction determined from computer simulations E. Belova and L. A. Chernozatonskii ………………………………….……..215 Oscillation of the charged doublewall carbon nanotube V. Lykah and E. S. Syrkin ……………………………………….…………...217 Polymer chains behavior in nanotubes: a Monte Carlo study K. Avramova and A. Milchev ……………………..………………………….219 Carbon nanotubes as ceramic matrix reinforcements C. Balázsi, F. Wéber, Z. Kövér, P. Arato , Z. Czigány, Z. Kónya, I. Kiricsi, Z. Kasztovszky, Z. Vértesy and L. P. Biró ……………………..……………..221 Carbon nanotubes as polymer building blocks F. M. Blighe, M. Ruether, R. Leahy and W. J. Blau …….…………………...223 Synthesis and characterization of epoxysinglewall carbon nanotube composites D. Vrbanic, M. Marinsek, S. Pejovnik, A. Anzlovar, P. Umek and D. Mihailovic ..…………………………………………………………..225 Vapour grown carbon nanofibers – polypropylene composites and their properties V. Chirila, G. Marginean, W. Brandl and T. Iclanzan ....................................227 Part VI. Applications Nanotechnology: challenges of convergence, heterogeneity and hierarchical integration A. Vaseashta …………………………………………………………………229 Behavior of carbon nanotubes in biological systems D. G. Kolomiyets …………………………………………………………….231 Molecular dynamics of carbon nanotubepolypeptide complexes at the biomembranewater interface K. V. Shaitan, Y. V. Tourleigh and D. N. Golik ……………………….….….233 Thermal conductivity enhancement of nanofluids A. Cherkasova and J. Shan …………………………………………………..235 Carbon nanotubes as advanced lubricant additives F. Dassenoy, L. JolyPottuz, J. M. Martin and T. Mieno ……………………237
ix Synthesis and characterization of iron nanostructures inside porous zeolites and their applications in water treatment technologies M. Vaclavikova, M. Matik, S. Jakabsky, S. Hredzak and G. Gallios ….…….239 Nanostructured carbon growth by an expanding radiofrequency plasma jet S. I. Vizireanu, B. Mitu, R. Birjega, G. Dinescu and V. Teodorescu ………...241 Design and relative stability of multicomponent nanowires T. Dumitric, V. Barone, M. Hu and B. I. Yakobson .......................................243 Modeling of molecular orbital and solid state packing polymer calculations on the bipolaron nature of conducting sensor poly (pphenylene) I. Rabias, P. Dallas and D. Niarchos ………………………………………..245 Nd:LSB microchip laser as a promising instrument for Raman spectroscopy V. Parfenov …………………………………………………………………..247 Subject Index…...……..…………………………………………………… 249 Author Index……..………………………………………………………… 251
PREFACE
It is about 15 years that the carbon nanotubes have been discovered by Sumio Iijima in a transmission electron microscope. Since that time, these long hollow cylindrical carbon molecules have revealed being remarkable nanostructures for several aspects. They are composed of just one element, Carbon, and are easily produced by several techniques. A nanotube can bend easily but still is very robust. The nanotubes can be manipulated and contacted to external electrodes. Their diameter is in the nanometer range, whereas their length may exceed several micrometers, if not several millimeters. In diameter, the nanotubes behave like molecules with quantized energy levels, while in length, they behave like a crystal with a continuous distribution of momenta. Depending on its exact atomic structure, a singlewall nanotube –that is to say a nanotube composed of just one rolledup graphene sheet– may be either a metal or a semiconductor. The nanotubes can carry a large electric current, they are also good thermal conductors. It is not surprising, then, that many applications have been proposed for the nanotubes. At the time of writing, one of their most promising applications is their ability to emit electrons when subjected to an external electric field. Carbon nanotubes can do so in normal vacuum conditions with a reasonable voltage threshold, which make them suitable for coldcathode devices. Nanotubes are also good candidates for the design of composite materials. They can increase the conductivity, either electrical or thermal, of polymer matrices which they are embedded in at a few weight percents, while improving the mechanical resistance of the materials. Most spectacular, but still far from industrialization, is the nanotubebased fieldeffect transistor. Here, a singlewall semiconducting nanotube, contacted to two electrodes, may block or may transmit an electric current depending on the potential applied to a gate electrode placed at near proximity. Many other applications are foreseen, among which nanoscopic gas sensing in which one property of the nanotube, sensitive to adsorbed molecules, is measured. Gas selectivity may be realized by a suitable functionalization of the nanotubes. Optical and optoelectronic properties of singlewall nanotubes are also promising for infrared applications. While the list of potential applications increases every month, the basic properties of intrinsic nanotubes are well documented and relatively well understood. Only relatively, because there remain several important open issues. Manybody effects, although predicted to occur in onedimensional systems since a long time, are not clearly evidenced. Luttingerliquid behavior, xi
xii for instance, is not fully recognized by experiments on metallic nanotubes. Excitons in semiconducting tubes constitute another topic of recent, sometimes controversial debates. More important, perhaps, the synthesis and growth mechanisms of the carbon nanotubes are not clearly pinned out. It is remarkable that these beautiful molecules can be produced in such many different physical and chemical conditions (electric arc discharge, catalytic chemical vapor deposition, laser ablation ...). Partly due to that, it is still not possible at the time of writing to produce nanotubes with all the same structure in a controllable way. Largescale, but detailed characterization of the nanotubes, like with any other nanostructures, remains a great experimental challenge that will need to be overcame. Whether or not nanotubes will have important industrial applications is not the essential point for the time being. What can be given for sure is that the carbon nanotubes have triggered an intense research activity thanks to which nanotechnology is developing so fast. The nanotubes are indeed ideal objects to deal with in this context before other nanostructures, perhaps, will supplement them and will open the way to real technological applications. In this book, many aspects of the nanotubes are either touched or described in details. The book is a snapshot, incomplete perhaps, of the state of the art at the time where the ASI took place, on the shore of the Black Sea. We gratefully acknowledge the generous support from the NATO Scientific and Environmental Affairs Division and the University of Namur. We thank all authors for preparing highquality manuscripts. V. N. Popov Sofia Bulgaria November 2005
Ph. Lambin Namur Belgium
ORGANIZING COMMITTEE
CoDirector Prof. Philippe Lambin Département de Physique Facultés Universitaires NotreDame de la Paix Namur, BELGIUM CoDirector Prof. Valentin Popov Faculty of Physics University of Sofia Sofia, BULGARIA Scientific Chairman Prof. Hans Kuzmany Universität Wien Institut für Materialphysik Wien, AUSTRIA Scientific Advisor Prof. Angel Rubio Dpto. Fisica de Materiales Facultad de Quimicas U. Pais Vasco San Sebastian/Donostia, SPAIN Scientific Advisor Prof. Minko Balkanski Université Pierre et Marie Curie Paris, FRANCE
xiii
Part I. Synthesis and structural characterization
ARC DISCHARGE AND LASER ABLATION SYNTHESIS OF SINGLEWALLED CARBON NANOTUBES BJÖRN HORNBOSTEL,* MIRO HALUSKA, JIRKA CECH, URSULA DETTLAFF, SIEGMAR ROTH MaxPlanckInstitute for SolidState Research, Stuttgart, Germany
Abstract. The laser ablation synthesis of carbon nanotubes is contrasted with the arc discharge method with respect to the synthesis product. A novel combination of two laser systems of different wavelengths for the laser ablation synthesis is presented. The impact of sulfur on the synthesis process is discussed. An excerpt of our quality control protocol is presented.
Keywords: singlewalled carbon nanotubes; laser ablation; arc discharge; synthesis; CO2; Nd:YAG; sulfur; quality control protocol
1. Introduction The prospects for a wide range of applications of singlewalled carbon nanotubes (SWCNT, SWNT) rely on the development of a costeffective largescale production. The three main roots for SWNT synthesis are laser ablation (LA), arcdischarge (arcD)13 in a Krätschmer4 reactor, and chemical vapor deposition (CVD). Besides these, there are plenty of scions where different methods merge into each other, other ambient conditions are used (e.g., submerged arcD58) or where other supporting energy sources sustain (e.g. PECVD912). Here, only the laser ablation and the standard Krätschmer arc discharge method are regarded. These two synthesis methods in plasma aggregate state rely on the presence of a catalyst for the production of SWNT. Historically, the arcD method was the first technique for multiwalled carbon nanotubes1 (MWCNT, MWNT) and singlewalled carbon nanotubes.2,3 By this method one is able to produce roughly 100 mg/min of SWNTcontaining soot. However, due to highly fluctuating conditions in the plasma *To whom correspondence should be addressed. Björn Hornbostel; email: [email protected]
1 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 1–18. © 2006 Springer. Printed in the Netherlands.
2 plume of the light arc it is difficult to keep a favorable condition for a long period. There are attempts to keep the conditions stable, which is essential for the production of flawless nanotubes at a high yield. Biggest drawback of the Krätschmer method is the relative high amount of undesired byproducts, as fullerenes, graphite and amorphous carbon (aC). In 1995 Guo et al.13 synthesized nanotubes by laser ablation for the first time. The usage of concentrated light directed on a target to evaporate material and to create highly reactive plasma is a mean to engender more controlled conditions in the plume. In the following we will present our way to synthesize SWCNT by ArcD and LA. Therefore, we will recapitulate the state of the art and our own experiences with both methods. Firstly, we will elaborate on the technical equipment and parameters. Later we will describe briefly how we evaluate asproduced and purified SWCNT material by our quality control protocol. 2. Technology and experimental 2.1. DC ARC DISCHARGE TECHNOLOGY
Between a pair of graphite electrodes in an inert gas atmosphere a DC electric arc discharge is ignited by e.g. a short contact between both. As a result, electrons exit from the cathode and form an electron cloud. At this time the voltage is still zero. As dissociating both from each other the resulting empty space between them is filled up by electrons and by the ambient gas. The electrons are accelerated towards the anode and on their way they ionize the gas molecules cascadelike by impact ionization. Positive charge carriers move to the negative counter pole. As soon as enough charge carriers are situated in the conducting channel, the arc ignites. At the moment of electrode separation the voltage source has to deliver full voltage. The incoming electrons give up their kinetic energy to the anode which causes the material to sublimate. The cathode will be cooled by discharge work of the electrons. The electrodes are typically watercooled graphite rods separated by few mm. A bias voltage of 15 to 35V is applied at currents between 50 and 120A. Achieving stable discharge plasma is the main factor in generating an environment favorable to nanotube growth. This is not so easy, as the anode is consumed and therefore has to be tracked towards the cathode continuously. Furthermore, the stability of an electric arc is limited due to its moving nature on the cathode and anode surface. Additionally, after already a few minutes the resulting uneven consumption of the anode and buildup of material on the cathode side causes a further instability in the DC arc. The rate of synthesis of a lab Krätschmer reactor can surpass 100 mg/min.
3
Figure 1. Schematic of a standard DC Krätschmer reactor (arc discharge).
A vast amount of studies on the fundamental technical parameters to optimize the arcD method has been accomplished and people are still working on it. Apparently, the yield and the properties of the nanotubes are not just dependent on the anode composition, the background gas pressure, the gas composition, ... but also linked to the apparatus size and its geometry, the thermal gradients and other influences of the sublimation system. Due to sometimes total different results (yields) at same parameter sets in different reactors it is hard to deduce a general rule for an optimized production. 2.2. ARC DISCHARGE PARAMETERS
The catalysts can be introduced by either drilling a hole into the anode or filling it up with a catalystcarbon mixure or by intermixing a larger portion of C and catalyst and then pressing to a rod. As for the laser ablation (LA) it was found that using bimaterial plus C mixtures of Co, Ni, Y, and Fe favors the production of larger quantities of SWNT. Furthermore, people agree upon Co/Y and Ni/Y intermixtures being the most effcient. Recently, Itkis et al.14 published their results on optimum anode compositions. Helium at around 500800 mbar is most favorable for the SWCNTproduction. Additional gases (e.g. H2) or substances (e.g. S) are able to influence the yield, the diameter distribution and the quality of the product. We currently exploit the advantage of sulfur in the process.
4 In the arc discharge production method sulfur functions as a SWNTs growth promoter and surfactant when added together with Ni/Fe/Co,15 Ni/Co,15,16 Ni/Y/Fe or Ni/Ce/Fe17 catalysts into the anode. Sulfur is used as promoter in other nanotube production methods as well, like in the solar energy evaporation method18 with Ni/Co and in the laser vaporization method19 with Ni/Co/Fe. It was found that sulfur alone without metal catalysts does not catalyze the SWNTs growth process.20 Typically for all methods mentioned and all catalysts used, the addition of sulfur increases the yield of nanotubes and broadens the tube diameter distribution to extend out to 6 nm. Exception is the formation of nanotubes with small diameters in the range of 0.91.1 nm as reported by Alvarez et al.18 Our work is focused on the influence of the S concentration added to Fe/Y catalysts on the yield of nanotubes and their properties. We chose Fe instead of Ni because it is much easier to remove it from the arc product. Our primary aim is to find satisfactorily high yields of high quality SWNT web material. One of the growth controlling parameters is the distance of the electrodes during the arc process. It influences the evaporation rate (temperature of anode) and the yield of web product. The highest the yield was obtained for the distance ~ 2 mm. The anode evaporation rate was 1.2 g/min.
Figure 2. The yield and the evaporation rate dependences on the electrode gap distance.
As it was discussed in the above mentioned publications sulfur does not dissolve in the bulk of the transition metals but adsorbs on the surface. The metalsulfur interactions change surface tension and melting point of small droplets of metals. This can support the creation of SWNTs for metals which in pure form catalyze badly. The overcritical concentration of S has a poisoning effect on SWNTs growth similar as for many other catalytically controlled processes. The highest yield of web product containing the smallest
5 concentration of metals was obtained for the sample C where the composition of the anode is Fe:Y:S:C 6.6 at.%:1.1 at.%:1.6 at.%:90.7 at.%. A more detailed overview on this work is presented in.21
Figure 3. The yield dependence over the concentration of sulfur.
Figure 4. The relative purity determined by NIRspectroscopy.13,14 A dashed line depicts the trend. Sulfur was used to attain better results. The background pressure was 550mbar.
2.3. LASER ABLATION TECHNOLOGY
The standard SWNT growth setup consists of a quartz tube (~25 mm diameter, 10001500mm length) mounted inside a hinged tube furnace that can operate at a temperature of 1200°C. The quartz tube is sealed to vacuum components. The laser beam enters the quartz tube through a Brewster window, which should be
6 plated by some antireflex layer for the incoming beam. Inert gas, e.g., Argon, or mixed gas compositions are introduced at the upstream side of the tube. The gas feeding is controlled by a mass flow controller and the pressure by a preassure controller downstream. Before the gas exits the system it passes a watercooled brass collector and a filter to collect the SWNTs. The brass collector is inserted into the quartz tube and positioned just outside the furnace. A rotating rod is led through the watercooled collector. A target consisting of carbon and metall catalysts is attached to it. This is to ensure a more homogenious ablation of the target. In addition it is appropriate to have the laser beam scanning over the target. The carrier gasflow sweeps most of the carbon species produced by the laser evaporation out of the furnace zone depositing it as soot on a watercooled copper rod. Usually the ablation laser is a Nd:YAG opperating at 1064 nm or 532 nm, respectively. Specific values of those Nd:YAG systems are between 300 mJ and 1.5 J per pulse at Co >> Fe > Pt, where yields decrease by a few percent in the same order. The synthesized tubes are few, isolated and often covered by aC. In case of a bimaterial combination a weblike structure is produced where thick and relatively clean bundels of SWNT can be observed. Yields may be over a hundred times higher than achieved with onedopant targets. Guo et al. found the following sequence for bimaterial compositions and Nd:YAG systems: Ni/Co ~ Co/Pt > Ni/Pt >>Co/Cu. Munoz et al.24 examined bimaterial mixtures for CO2laser systems and concluded Ni/Co~Ni/Y>>Ni/Fe~Ni/La~Co/Ni. For our initial experiments we use mainly Ni/Y (4.2at.%/1at.%), Ni/Y (2at.%/0.5at.%) and Ni/Co (2at.%/2at.%) plus carbon compositions. One target is around 8g in weight. The very dryed mixture is mixed by a toothpick first before we put it into a tumbling machine for at at least 48h to assume a macroscopic homogenious intermixture. Afterwards a 15mmdiameter pellet is pressed in a 150°C warm tool at 8kN. The optimum operation parameters of our laserfurnace setup, concerning preassure, carrier gas, flow and temperature as well as applied laser power will be puplished elsewhere as soon as our experiments are completed. Our momentary production record at a quite acceptaple quality is at 200mg/h without S. With only Nd:YAG we attchieved only 60mg/h. This is a triplication of the production rate. The composition of the target was C:Ni:Co 96 at.%: 2 at.% : 2 at.%. The pure Ar gas flow was controlled to be 350 sccm. 3. Synthesis product evaluation Our experience is that a highly optimized process for arcD may yield up to 55 vol.% SWNT. The ArcD process delivers in contrast to LA rather short carbon nanotubes (some microns) with a quite broad diameter distribution. Laser ablated SWNTs are remarkably uniform in diameter and selforganized into bundles via vanderWaals forces. Consisting of 100 to 500 SWNTs, their diameters range from 10 to 20 nm and their lengths can reach values between 10 and some 100 Pm. In the synthesis of arcD tubes quite low defect densities can be observed. The high temperatures in the plasma usually ensure a complete closure of the tube lattices. An even lower defect density can be expected in
9 singlewalled nanotubes produced in the laser ablation setup. Here, the few and minor fluctuations in the process conditions are less important. The more stable conditions are also the reason for the longer, more identical tubes and for the higher purity. The energy in the arcD process is high enough to evaporate the material carefully. But due to the alternating conditions a favorable condensation of the material is hindered. By TEM one can observe structural differences between both synthesis products concerning the catalysts: ArcD tubes usually encase the catalyst particle with which help the tube grows. In LA material it is more common to have the particle somewhere in the soot but not encased or attached to the tube. For getting rid of the catalyst content in the endproduct more easily this cartelistic is opportune. The production rate in the arcD is by far higher. While the rate of an arcD process is usually measured on the 100 mg/min scale the production rate in the LA is expressed by the 10mg/h or 100mg/h scale. But one has to keep in mind that the yield of SWNT in the as produced arcD soot does not surpass 55 vol.% today. Of course, today it is not possible to produce SWNT of LA quality by ArcD. There are possibilities to clean the soot from all unwanted content, like catalysts, aC, fullerenes and graphite by purification. However, according to our experiences so far, this might be applicable for small quantities or for aC and fullerenes but in case of an overall purification and a mass production these cleaning steps for arcD material would turn out to be the most cost intensive. Furthermore, by running through certain purification steps, not just the unwanted material will be attacked but also the SWNTs will be harmed to some extent. Quality characterization in the SWNT field is a complicated issue. To ensure the comparability of our measurements we have been working on a protocol for quality control of singlewalled carbon nanotube material. This protocol is regularly updated.30 The momentary version of the protocol includes standards for homogenization, bucky paper preparation, composites, electrical, xray diffraction SEM, TEM, optical spectroscopy and Raman measurements. We stress the importance of working with large homogenized batches (Presently, we use 100 g batches of ArcD material, whenever possible). A short excerpt of our protocol is presented here. 3.1. ELECTRICAL MEASUREMENTS
Figure 6 shows a simple device to measure the conductivity of bucky papers and of thin composite films using the 4 leads method. A strip punched off from a bucky paper or a composite film is placed under the two outer (current) electrodes and held with 5 mm wide clamps. Two inner (voltage) electrodes
10 with 10mm distance are put onto the sample orthogonally to the surface and pressed with a constant force of 0.6 N. To this end a weight made of a stainless steal block is placed on the body holding and separating the measuring electrodes (insulating PTFE/PE, 1018 :cm).
Figure 6. Device to measure the electric conductivity with the fourleadsmethod.
Measurement: Constant current (mA) is applied to determine the voltage drop between the inner electrodes. The thickness of the strip (usually between 60 and 100 µm) is measured. At least three strips of one bucky paper should be measured to estimate the mean electrical conductivity and the standard deviation. The conductivity1 is calculated as follows: R = ǻU / I ı = 1/R ǜ L/(T*W) L = 10 mm W = 2 or 5 mm T
§ conductivity in Siemens per centimeter [S/cm = (ȍcm)1] § distance between the inner electrodes (constant) § width of the sample (constant) § lowest thickness of the bucky paper strip
______ 1
Remark: A strong scattering of the conductivity values (larger than 15%) is evidence of problems in the preparation of bucky paper and the preparation should be repeated.
11 3.2. ELECTRON MICROSCOPY
Note that electron microscopy (SEM and TEM) is not well suited to characterise large batches. Sample preparation and image evaluation is very "selective", i.e. the result depends on the skill and patience of the spectroscopist and on what he hopes to see. Usually "best" data are shown rather than "representative" data. But the methods are very useful to check whether in a particular sample "there are nanotubes at all". When new synthetic routs for nanotube production are developed, SEM very often gives the first positive hint of success. SEM characterization can be carried out on powder deposited on the sticky side of scotch tape or on a bucky paper.
Figure 7. Visual comparison of SWNTmaterial produced by ArcD (left) and LA (right). Both TEMpictures were taken at same magnification. In the ArcD material dirt and irregular large particles are clearly recognizable.
3.3. XRAY POWDER DIFFRACTION
We execute xray measurements in two ways. The first one is to use a Lindeman tube. Here, the nanotube powder is filled into a Lindeman tube (thin glass capillary, as commonly used for Xray powder diffraction), diameter 0.3 or 0.5 mm. The second possibility is to use a flat bed sample holder. About 5 mg of powder is distributed on a scotch tape and put into the sample holder. Both sample holders rotate while measuring. Fig. 8 (left) shows a typical powder diffractogram of carbon arc material after various purification steps (Copper KD radiation, 20 minutes exposure time, Lindeman tube, multidetector). If there are nanotube bundles in the sample, a peak will show up at 6°, which originates from the triangular lattice of the tubes in the bundle (lattice constant about 10 Å). The central curve shows a slight indication of such a peak. The peak at ~26° corresponds to graphite. At higher scattering angles there are the peaks of metal (raw material) or metal oxides (after heating
12 in air). For quantitative measurements the SWNT material is mixed in a ratio of 1:1 with fullerene C60. Xray powder diffraction patterns are recorded using the flat bed method (By filling into a Lindeman tube the sample would demix). The area of the graphite peak at 26° is compared with that of the second fullerene peak at 16° (which is used as a marker).
1
1 2 3
2 3
Figure 8. A typical powder diffractogram of ArcD material after various purification steps.
Figure 9. The pattern of a sample with unknown graphite concentration is shown. Now the peak area between graphite and fullerene marker compare like 1 to 5, indicating that the sample contains about 20% of graphite contamination (In addition to crystalline graphite there probably also is an important contamination of turbostratic graphite leading to the broad peak between 15° and 25°).
13 3.4. OPTICAL SPECTROSCOPY
For a first orientation, absorption spectra of films sprayed onto a glass substrate are useful. Therefore, SWNTs are suspended in 1% SDS aqueous solution and sprayed onto microscope slides using the airbrush technique. Fig. 9 shows the spectrum of SWCNT after subtraction of an appropriate baseline. The peaks are assigned to transitions between the van Hove spikes in the electronic density of states of nanotubes. S11 and S22 are related to semiconducting tubes and M11 originates from metallic tubes. S11 depends very much on the chemical environment of the sample (e.g. doping) and is not well suited for a quantitative analysis. The background is due to carbonaceous byproducts, like amorphous carbon and graphitic particles, and also to excitations in the nanotubes themselves (but not directly related to the van Hove spikes). For a quantitative analysis we follow Itkis et al. method.30,31 This analysis only yields relative nanotube concentrations, relative to a standard of known concentration. For ArcD material we use the standard (A(S22)/A(T) = 0.141) published in.31 For LA material we use A(S22)/A(T) = 0.211, a value which we measured on some material which was delivered by FZ Karlsruhe (Germany) and purified by ourselves. NIR spectroscopy and electrical measurements in bucky paper correlate well as can be seen in Figure 10.
Figure 10. Spectrum of SWCNT after the subtraction of a proper baseline.
14
Figure 11. NIR spectroscopy in solution and electrical measurements in bucky paper correlate well. Here we purified ArcD material in a hot furnace at a set temperature for differently long periods.
3.5. RAMAN SPECTROSCOPY
Figure 12 shows a Raman spectrum measured by a Raman microscope, excitation line O = 632 nm. The double line at 1600 cm1 is the so called Gline coming from the stretching of conjugated double bonds. The lines around 200 cm1 are from the radial breathing mode (RBM) and depend on the nanotube diameter, the line at 1300 cm1 (Dline) is from the multiple phonon scattering and from the nanotubes defects. Spectra are normalized to the D*mode (2600 cm1). Note that Raman spectra are resonance enhanced and depend very much on the excitation line. For our believe, Raman spectra are not useful for quantitative analysis, but the intensity of the Dline is often taken as evidence of many defects on the tube or of the presence of amorphous carbon in the sample. In addition, like SEM and TEM images, Raman spectra often give the first hint of the presence of singlewalled carbon nanotubes in a sample and then justify further investigations.
15
Figure 12. Raman lines from ArcD and LA material at 632.817nm. LA nanotubes have a narrower diameter distribution.
4. Summary In comparison to the laser ablation method, the arc discharge SWCNT synthesis is still the most efficient method to produce large quantities of nanotubes. Additionally, it is also the most widly used one which also traces back to the relative simple basic setup. In combination of optimized process parameters and additional ingredients like sulfur yield and quality can be influenced positively. Although the relative amount of SWCNT in the soot can be increased by varying the process parameters and by gas/ingredient variations the synthesis process regime is difficult to control. Hence, in case of need for very pure material with a narrow diameter distribution the laser ablation method is the process of choice. Our approach to use a Nd:YAG system at 1064nm and a CO2 system at 10.6 microns simultaneously is already capable to triple the production rate to 200mg/h. A main question is whether laser ablation may ever be a synthesis method to produce bulk quantities or if it stays a method for nanotubes for specialized applications. For the evaluation of SWNT material we developed a protocol34 which is regularly updated. It is one of the most important means for our work on synthesis and purification.
16 References 1. 2. 3.
4. 5. 6.
7. 8.
9.
10.
11.
12.
13. 14.
15.
16.
17.
S. Iijima, Helical microtubules of graphitic carbon, Nature 354, 56–58 (1991). S. Iijima and T. Ichihashi, Singleshell carbon nanotubes of 1nm diameter, Nature 363, 603–605 (1993). D. S. Bethune, C. H. Kiang, M. S. de Vries, G. Gorman, R. Savoy, J. Vazquez, and R. Beyers, Cobaltcatalysed growth of carbon nanotubes with singleatomiclayer walls, Nature 363, 605–607 (1993). W. Krätschmer, L. O. Lamb, K. Fostiropoulos, and D. R. Huffman, Solid C60: a new form of carbon, Nature 347, 354–358 (1990). Y. L. Hsin, K. C. Hwang, F.R. Chen, and J.J. Kai, Production and insitu metal filling of carbon nanotubes in water, Advanced Materials 13(11), 830833 (2001). H. W. Zhu, X. S. Li, B. Jiang, C. L. Xu, Y. F. Zhu, D. H. Wu and X. H. Chen Formation of carbon nanotubes in water by the electricarc technique. Chem. Phys. Lett. 366, 664669 (2002). H. Lange, M. Sioda, A. Huczko, Y. Q. Zhu, H. W. Kroto and D. R. M. Walton, Nanaocarbon production by arc discharge in water, Carbon 41, 16171623 (2003). L. P. Biró, Z. E. Horváth, L. Szalmás, K. Kertész, F. Wéber, G. Juhász, G. Radnóczi and J. Gyulai, Continuous carbon nanotube production in underwater AC electric arc, Chemical Physics Letters 372, 399402 (2003). Ch. Bower, O. Zhou, W. Zhu, D. J. Werder, and S. Jin, Nucleation and growth of carbon nanotubes by microwave plasma chemical vapor deposition, Applied Physics Letters 77, 27672769 (2000). M. Chhowalla, K. B. K. Teo, C. Ducati, N. L. Rupesinghe, G. A. J. Amaratunga, A. C. Ferrari, D. Roy, J. Robertson, and W. I. Milne, Growth Process Conditions of Vertically Aligned Carbon Nanotubes Using Plasma Enhanced Chemical Vapor Deposition, J. Appl. Phys. 90(10), 53085317 (2001). L. Delzeit, C. V. Nguyen, R. M. Stevens, J. Han, and M. Meyyappan, Growth of carbon nanotubes by thermal and plasma chemical vapour deposition processes and applications in microscopy, Nanotechnology 13, 280284 (2002). J. Han, J.B. Yoo, C. Y. Park, H.J. Kim, G. S. Park, M. Yang, I. T. Han, N. Lee, W. Yi, S. G. Yu, and J. M. Kim, Tip Growth Model of Carbon Tubules Grown on the Glass Substrate by Plasma Enhanced Chemical Vapor Deposition, J. Appl. Phys. 91, 483486 (2002). T. Guo, P. Nikolaev, A. Thess, D. T. Colbert, and R. E. Smalley, Catalytic growth of singlewalled nanotubes by laser vaporization, Chem. Phys. Lett. 243, 49–54 (1995). M. E. Itkis, D. E. Perea, S. Niyogi, J. Love, J. Tang, A. Yu, C. Kang, R. Jung, and R. C. Haddon, Optimization of the NiY Catalyst Composition in Bulk Electric Arc Synthesis of SingleWalled Carbon Nanotubes by Use of NearInfrared Spectroscopy, J. Phys. Chem. B 108, 1277012775 (2004). Y. S. Park, K. S. Kim, H. J. Jeong, W. S. Kim, J. M. Moon, K. H. An, D. J. Bae, Y. S. Lee, G.S. Park, and Y. H. Lee, Low pressure synthesis of singlewalled carbon nanotubes by arc discharge, Synthetic Metals 126, 245251 (2002). C. Liu, H. M. Cheng, H. T. Cong, F. Li, G. Su, B. L. Zhou, M. S. Dresselhaus. Synthesis of macroscopically long ropes of wellaligned singlewalled carbon nanotubes, Advanced Materials 12, 11901192 (2000). H. W. Zhu, B. Jiang, C. L. Xu, and D. H. Wu, Synthesis of High Quality Singlewalled Carbon Nanotube Silks by the Arcdischarge Technique, J. Phys. Chem. B 107, 65146518 (2003).
17 18. L. Alvarez, T. Guillard, J. L. Sauvajol, G. Flamant, and D. Laplaze, Solar production of singlewall carbon nanotubes: growth mechanisms studied by electron microscopy and Raman spectroscopy, Appl. Phys. A 70(2), 169–173 (2000); L. Alvarez L., Guillard T., Sauvajol, G. Flamant, and D. Laplaze, Growth mechanisms and diameter evolution of single wall carbon nanotube, Chem. Phys. Lett. 342, 714 (2001). 19. S. Lebedkin, P. Schweiss, B. Renker, S. Malik, F. Hennrich, M. Neumaier, C. Stoermer, and M. M. Kappes, Singlewall Carbon Nanotubes with Diameters Approaching 6 nm Obtained by Laser Vaporization, Carbon 40, 417 423 (2002). 20. C.H. Kiang, W. A. Goddard III, R. Beyers, J. R. Salem, D. S. Bethune, Catalytic Synthesis of SingleLayer Carbon Nanotubes with a Wide Range of Diameters, J. Phys. Chem. 98, 66126618 (1994). 21. M. Haluska, V. Skakalova, D. L. Carroll, and Z. Roth, The Influence of Sulfur Promoter on the Production of SWCNTs by the ArcDischarge Process, in: Electronic Properties of Novel Materials, H. Kuzmany, J. Fink, M. Mehring, S. Roth (eds.), American Institute of Physics, New York, 2005, AIP Conference Proceedings, in print. 22. W. K. Maser, E. Munoz, M. T. Martınez, A. M. Benito, and G. F. de la Fuente, Study of parameters important for the growth of single wall carbon nanotubes, Opt. Mater. 17, 331– 334 (2001). 23. E. Munoz, W. K. Maser, A. M. Benito, M. T. Martınez, G. F. de la Fuente, Y. Maniette, A. Righi, E. Anglaret, and J. L. Sauvajol, Gas and pressure effects on the production of singlewalled carbon nanotubes by laser ablation, Carbon 38, 1445–1451 (2000). 24. E. Munoz, W. K. Maser, A. M. Benito, M. T. Martınez, G. F. de la Fuente,A. Righi, E. Anglaret, and J. L. Sauvajol, The influence of the target composition in the structural characteristics of single walled carbon nanotubes produced by laser ablation, Synth. Metals 121, 1193–1194 (2001).. 25. A. P. Bolshakov, S. A. Uglov, A. V. Saveliev, V. I. Konov, A. A. Gorbunov, W. Pompe, and A. Graff, A novel CW laserpowder method of carbon singlewall nanotubes production, Diam. Rel. Mater. 11, 927–930 (2002). 26. W. K. Maser, A. M. Benito, E. Muñoz, G. M. de Val, M. T. Martínez, Á. Larrea, and G. F. de la Fuente, Production of carbon nanotubes by CO2laser evaporation of various carbonaceous feedstock materials, Nanotechnology 12, 147151 (2001). 27. A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. H. Lee, S. G. Kim, A. G. Rinzler, D. T. Colbert, G. E. Scuseria, D. Tomanek, J. E. Fischer, and R. E. Smalley, Crystalline ropes of metallic carbon nanotubes, Science 273, 483–487 (1996). 28. A. G. Rinzler, J. Liu, H. Dai, P. Nikolaev, C. B. Huffman, F. J. RodrıguezMacıas, P. J. Boul, A. H. Lu, D. Heymann, D. T. Colbert, R. S. Lee, J. E. Fischer, A. M. Rao, P. C. Eklund, and R. E. Smalley, Largescale purification of singlewall carbon nanotubes: process, product, and characterization, Appl. Phys. A 67, 29–37 (1998). 29. Ch. T. Kingston, Z. J. Jakubek, S. Denommee, and B. Simard, Efficient laser synthesis of singlewalled carbon nanotubes through laser heating of the condensing vaporization plume, Carbon 42, 1657–1664 (2004). 30. U. DettlaffWeglikowska, J. Wang, J. Liang, B. Hornbostel, and S. Roth, Purity Evaluation of Bulk Single Wall Carbon Natotube Materials, in: Electronic Properties of Novel Materials, H. Kuzmany, J. Fink, M. Mehring, S. Roth (eds.), American Institute of Physics, New York, 2005, AIP Conference Proceedings, in print. 31. M. E. Itkis, D. Perea, S. Niyogi, S. Rickard, M. Hamon, H. Hu, B. Zhao, and R. C. Haddon, Purity Evaluation of AsPrepared SingleWalled Carbon Nanotube Soot by Use of Solution Phase NearIR Spectroscopy, Nano Lett. 3, 309314 (2003).
18 32. W. K. Maser, E. Munoz, A. M. Benito, M. T. Martınez, G. F. de la Fuente, Y. Maniette, E. Anglaret, and J.L. Sauvajol, Production of highdensity singlewalled nanotube material by a simple laserablation method, Chem. Phys. Lett. 292, 587–593 (1998). 33. F. Kokai, K. Takahashi, M. Yudasaka, and S. Iijima, Emission Imaging Spectroscopic and Shadow graphic Studies on the Growth Dynamics of Graphitic Carbon Particles Synthesized by CO2 Laser Vaporization, J. Phys. Chem. 103, 86868693 (1999). 34. This protocol is a guidance for sample characterization and quality control within the SPANG consortium. It is hoped that by applying this protocol, reproducible results will be obtained at the different laboratories. To facilitate cooperation with other consortia and quite general with colleagues worldwide, the protocol is also spread outside the EU project SPANG. The protocol is regularly updated. We appreciate any interest for the protocol and any feedback from colleagues working in the field.
SCANNING TUNNELING MICROSCOPY AND SPECTROSCOPY OF CARBON NANOTUBES LÁSZLÓ P. BIRÓ* Research Institute for Technical Physics and Materials Science, H1525, POB 49, Budapest, Hungary PHILIPPE LAMBIN Facultés Universitaires NotreDame de la Paix, B5000, Rue de Bruxelles 61, Namur, Belgium
Abstract. This paper briefly reviews scanning tunneling microscopy with emphasis on the particularities that have the most significant influence on the STM and STS characterization of both singlewall and multiwall carbon nanotubes. The experimental results obtained on these carbon nanotubes are surveyed together with computer modeling of the STM images of carbon nanotubes and the role of structural defects is discussed.
Keywords: scanning tunneling microscopy (STM); scanning tunneling spectroscopy (STS); carbon nanotubes; computer modeling; structural defects
1. Introduction Carbon nanotubes (CNTs) are perhaps one of the most exciting members of the family of carbon nanostructures. The very quick development of this family was initiated by the discovery of fullerene almost two decades ago1. The very promising properties of the CNTs2,3 brought singlewall carbon nanotubes (SWCNTs) and multiwall carbon nanotubes (MWCNTs) in the focus of the attention of researchers in various fields ranging from nanoelectronics to composite materials. The straight carbon nanotubes were discovered by Iijima in 1991 (Ref. [4]). Later on, other more complex, but tubular type structures built of layers of sp2 hybridized carbon, like Yjunctions, helical coils and *To whom correspondence should be addressed: László P. Biró, Nanotechnology Department, Research Institute for Technical Physics and Materials Science, H1525, POB 49, Budapest, Hungary; email: [email protected]
19 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 19–42. © 2006 Springer. Printed in the Netherlands.
20 multiple coils have been predicted theoretically58 and soon observed experimentally by transmission electron microscopy (TEM) and scanning tunneling microscopy (STM).9 13
Figure 1. Structural models of carbon nanotubes: a) varieties of SWCNTs depending on the wrapping of the graphene layer; b) a MWCNT composed of three concentric SWCNTs (with permission from http://www.photon.t.utokyo.ac.jp/~maruyama/agallery/agallery.html).
A SWCNT is constituted from a one atomic layer thick sheet with graphitelike structure (graphene sheet) rolled into a perfect cylinder (Fig. 1a), while MWCNTs are built of a set of graphene cylinders (which can be regarded as individual SWCNTs) with increasing diameter, concentrically placed in each other (Fig. 1b). The diameter and helicity of a defectfree SWCNT are uniquely characterized by the vector C = na1 + ma2 Ł (n,m) called the rolling (or wrapping) vector, that connects crystallographically equivalent sites on the twodimensional graphene sheet. Here a1 and a2 are the graphene lattice vectors, and n and m are integer numbers (see Fig. 2). The interlayer spacing of MWCNTs is consistent with the characteristic one for turbostratic graphite of 0.34 nm (Ref. [14]). The typical diameter values of SWCNTs are around 1 nm, while MWCNTs have diameters of a few tens of nanometers. Their structure confers the nanotubes with special properties, very closely determined by the way in which the CC bonds in the graphene sheet are oriented with respect the nanotube axis, and by the diameter of the tube2,3 (Fig. 1). As a typical example: depending on the way in which the graphene sheet is wrapped into a cylinder, SWCNTs may be metallic or semiconducting.2 One third of all possible SWCNTs shows metallic behavior, i.e., their electronic density of state (DOS) at the Fermi energy is different from zero, while two thirds exhibit semiconducting behavior with zero DOS around the Fermi energy. Normally,
21 MWCNTs are constituted from a random alternation of semiconducting and metallic layers.15 Due to the strong relation between the atomic structure and the electronic properties, atomic resolution topographic STM images and scanning tunneling spectroscopy (STS) are very useful in the investigation and in the understanding of physical properties of carbon nanotubes. Presently, the STM is the only tool able to give information about both the atomic arrangement and electronic structure of the very same nanotube.
Figure 2. Schematic drawing showing the wrapping of a (11,7) SWCNT and its topographic STM image (after Ref. [47]). a) wrapping of the graphene sheet using a (11,7) wrapping vector C, the wrapping angle ), the vector T defining the tube axis and the vector H defining the direction of a row of hexagons appearing as dark spots on the STM image, b) atomic resolution, topographic STM image showing the relation of vectors T and H.
2. Brief overview of STM The basic concept of the STM is simple (Fig. 3): an atomically sharp metallic tip is brought in the proximity of a flat conducting sample within a distance s of the order of a few tenths of a nanometer. As their behavior is governed by the laws of quantum mechanics, electrons may tunnel with equal probability from the tip to the surface and vice versa. When an external bias V is applied between the surface and the tip, preferential tunneling will be imposed in one of the directions (depending on the polarity of V), which will yield an average tunneling current It. For maintaining continuous tunneling, the tipsample distance s has to be regulated using a piezo actuator and a feedback loop. The tunneling current measured at a given moment is taken as input signal for the feedback loop and compared with a preset value. The actuator will increase or reduce the gap between sample and tip depending on whether the difference of measured value of It is larger or smaller than the preset value.
22 A comprehensive overview of the theoretical models used to describe the tunneling through a one dimensional potential barrier, the simplest case of tunneling, is given by Wiesendanger.16 Two major conclusions arise from the various models:
Figure 3. Principle of operation of the STM.
The transmission T of electrons through a one dimensional potential barrier depends exponentially on the width s of the barrier
T v e 2 Fs .
(1)
The decay rate F, characterizing the decrease of the probability of the electron to cross over the barrier, is 2S [2m(V0 E )]1/ 2
F
h
,
(2)
where Vo is the average height of the barrier, E is the energy of the tunneling electron. Independently of the exact shape of the barrier, the strong exponential dependence of T with the barrier width s and the square root of the effective barrier height, (Vo  E)1/2 is typical of tunneling. The exponential dependence makes that the tunneling channel will be very narrow, thus making possible the atomic resolution. The tunnel current flowing between the tip and the sample at a bias voltage V will be: eV
I (V ) v ³ U s ( E ) U t ( E eV ) T ( E ) dE , 0
(3)
23 where Us is the sample electronic DOS, Ut is the DOS of the tip and T the transmission coefficient of the barrier. From Eq. 3, one may observe that the STM offers the opportunity to acquire not only topographic information about the sample but also information regarding the DOS of the sample. If the tip DOS is flat and the transmission coefficient T may be taken as constant, then Eq. 3 reduces to: eV
I (V ) v ³ U s ( E )dE .
(4)
0
Then, the quantity dI/dV, named differential resistance, willbe proportional toUs(V), the sample electronic DOS. However, T may be regarded as constant only in the limit of small voltage variations but it gives deviation from this simple dependence at large bias values.
Figure 4. Schematic representation of the STMnanotubesubstrate system: a) the STM tip when tunneling directly into the substrate; b); the STM tip when tunneling through a carbon nanotube (showed in transversal cross section), the shaded circle stands fro the carbon nanotube; c) 3D presentation of the STM tip, nanotube, substrate system, dimensions are in nanometers. Note that the nanotube is not integral part of the substrate.
In reality, the STM measurement is quite different from the simple case of tunneling through a one dimensional potential barrier. First of all, in the case of a sharp tip, one has a threedimensional potential barrier instead of a onedimensional one. The shape of real potential barriers may strongly differ from the rectangular shape assumed in deducing Eqs. 1 and 2. A widely used model for the interpretation of experimental STM images is the model given by
24 Tersoff and Hamann.17 In their theory, the tip is treated as a single s orbital, a constant DOS is assumed for it, and the tunnel current is taken as being proportional to the sample DOS computed at the location of the apex of the STM tip at the Fermi energy and taking a vanishingly small bias. A more comprehensive and detailed treatment of tunneling is given in the book on STM edited by Wiesendanger and Güntherodt.18 In the practical STM instrumentation, the positioning and scanning of the STM tip is usually achieved by piezoelectric actuators. It is important to emphasize that the generated image opposite to other image generating techniques, like analog photography, is constituted of pixels (256 x 256, 512 x 512, etc.). Therefore, zooming in using software tools into a recorded image unavoidably is reducing the amount of "measured" information. An STM can operate in several regimes; the three most important ones are as follows: Topographic (constant current) imaging: the width of the STM gap is controlled by a feedback loop keeping the value of the tunneling current at a setpoint value selected by the operator. The image is generated from the values of the voltage applied to the piezoactuator to maintain the constant value of the tunneling current. Provided the electronic structure at the sample surface is homogeneous, the topographic profile of the surface will be generated. Currentvoltage spectroscopy, frequently called scanning tunneling spectroscopy (STS). The scanning and the feedback loop are switched off, the value of the tunneling gap is fixed, and the bias voltage is ramped from –U to +U, and the corresponding current variations are recorded. The function dI/dV gives information about the local DOS of the sample. Currentvoltage spectroscopy and imaging (CITS): the feedback loop is switched on and of. While it is on, the scanning is done and image information is acquired (as in topographic imaging), in each image point, after the position of the tip has stabilized, the feedback loop is switched off and the full STS spectrum is acquired followed by switching on the feedback loop and continuing scanning. To understand the results of the STM imaging of carbon nanotubes, it is useful to keep in mind a few particularities of the tunneling system specific to this case: In order to be able to image a nanoscopic object, like a carbon nanotube in an STM, the object has to be placed on an atomically flat conducting surface: the substrate. Highly oriented pyrolitic graphite (HOPG), or Au (111) terraces are the most frequently used substrates.19 When a carbon nanotube is imaged, as opposite to the case of tunneling into a bulk sample, the electrons have to cross two tunnel gaps: one between the STM tip and the nanotube, the other between the nanotube and its support20,21 as shown in Fig. 4b.
25 The electronic structure of the carbon nanotube and its support may be different. As it follows from Eq. 3, the different values of Us(support)(V), and Us(nanotube)(V), will yield different values of tunneling current, thus having influence on the height values measured from topographic images.22 When the STM tip comes very close to the nanotube or when switching from pure tunneling to point contact imaging regime occurs,23 the STM tip may induce the elastic deformation of the nanotube. By changing the imaging conditions like It or V, if the mechanical contact was not as strong as to damage the tip or the sample, the point contact imaging can be "switched" on and off24 (Fig. 5).
Figure 5. A STM image taken on a MWCNT in conditions under which switchover to point contact imaging occurs. a) low magnification topographic image It = 1.2 nA, V = 100 mV. Note the depression in the top part of the MWCNT; b) linecut along the line marked in panel a), the depression of 0.1 nm depth is clearly visible in the topmost part of the nanotube; c) an atomic resolution image taken on the same nanotube with the same STM tip using It = 1.0 nA and V = 100 mV: the typical structure of graphite is well observed, the three axes running through the tunneling current maxima make an angle of 60o; d) linecut along the line marked in c) by AB: the curvature of the MWCNT is well observed. Note the corrugation amplitude of 0.05 nm equal to half of that characteristic for HOPG.
Convolution effects, arising between the shape of a nanometric object and the shape of the STM tip, may have a significant influence on the recorded STM image. This phenomenon may cause a significant apparent lateral broadening of the imaged tube25 (see Fig. 6).
26 The apparent diameter D may be related to the apparent height h of the tube by the approximate relation D = 8Rh)1/2, where R is the curvature radius of the tip.23 The height of the topographic profile h may be regarded as being close to the geometric diameter of the tube and h is often used as a measure of d, the geometric diameter of the tube. But the measured apparent height depends on a few factors: the adsorption distance of the nanotube above the substrate; on the difference between the tunneling distances above the tube and over the substrate arising from their different electronic properties; and on the possible compression of the tube by the STM tip. The magnitude of these effects is not known with precision. Using the apparent height h of the nanotube seems to underestimate the actual diameter by as much as 0.2–0.5 nm26 which may cause a significant error for a SWCNT, while it may be tolerable for a MWCNT. A fitting procedure of the tunneling current has been proposed27 for the better determination of the diameter from the measured height.
Figure 6. Illustration of the distortion due to tipsample convolution effects for objects with different shapes. When the object becomes sharper than the tip, the object will generate the image of the tip.
Distortion effects may also arise from a “floppy” tip. Such a tip may be produced when the rigid STM tips picks up a highly flexible carbon nanotube from the sample.28 In this case the attached nanotube will act as the tip generating the image. Such a tip may cause in the topographic STM image reduced apparent height and dips on the sides of the imaged nanotube. 3. STM and STS of multiwall carbon nanotubes The nanotubes discovered by Iijima4 in 1991 were MWCNTs, Fig. 1b. STM was perhaps the best suited tool to characterize the topography and the electronic properties of these nanoobjects which at that time were produced in extremely small quantities. Topographic and voltage dependent STM imaging of arcgrown MWCNTs in a bundle was first reported by Zhang and Lieber.29 The bundles were found to be composed of a parallel arrangement of closely packed nanotubes with diameters on the order of 10 nm. The first STS
27 measurement on individual carbon nanotubes was reported by Olk and Heremans.30 The experiment was carried out in air on MWCNTs grown by the electric arc method. The nanotubes were transferred onto an Au substrate using a suspension in ethanol prepared by ultrasonication. Both semiconductor and metallic carbon nanotubes were found. The comparison of measured energy gap and experimental diameter values with gap values predicted by theory31,32 showed an increasing deviation with decreasing tube diameter. The possible sources of this deviation are: the unavoidable error introduced in the measured diameter by tip/sample convolution effects (Fig. 6), the more complex structure of the tunneling region than in the well known case of a flat homogeneous sample20,22 (Fig. 4) and effects arising from the multiple walls of the MWCNTs (Fig. 1b). The role that the inner layers of a MWCNT may play during STM imaging is not fully clarified yet. Theoretical calculations show that the smaller inner layers of a MWCNT may rotate in the larger diameter outer layer.33 However, the first atomic resolution STM image was reported by Ge and Sattler34 on MWCNTs. Disregarding the curvature of the imaged object, a similar atomic resolution STM image was obtained as in the case of flat HOPG. The triangular lattice characteristic for HOPG (Fig. 7b) is caused by the nonequivalence of the carbon atoms (in the Bernal graphite stacking) due to interlayer interaction.35, 36 Thus a similar atomic resolution pattern in the case of a MWCNT like that of HOPG is an indication of an interlayer arrangement resembling that of HOPG. Additionally Ge and Sattler34 found a periodicity of 16 nm along the tube axis superimposed on the atomic lattice. This was interpreted as arising from the misorientation of the outer two layers of the MWCNT, analogous to the generation of Moiré patterns known in geometric optic. The STM observation of Moiré patterns on HOPG37 is well known experimentally. Despite that MWCNTs were discovered first, both their experimental and theoretic understanding is behind that of SWCNTs. Atomic resolution images acquired on MWCNTs grown by CVD22 were interpreted in a similar way like in the work by Ge and Sattler.34 The comparison of atomic resolution images of SWCNTs, MWCNTs and HOPG, taken within the same series of experiments, showed that the atomic resolution images on MWCNTs and HOPG are similar.38 Moreover, the difference between D and E sites (E sites do not have a neighbor in the layer below) was evidenced by STS too.38 Most of the atomic resolution STM images on MWCNTs have been obtained on large diameter tubes while on small diameter tubes only lowquality atomic resolution was reported,22 The latter may be related to the difficulty to achieve a HOPGlike stacking of the layers at small diameters.
28 The particularities in the DOS in the region of closed MWCNT ends were investigated using spatially resolved scanning tunneling spectroscopy and tightbinding calculations. Sharp resonant valence band states were found at the tube ends dominating the valence band edge and filling the band gap. Calculations show that the strength and position of these states with respect to the Fermi level depend sensitively on the relative positions of pentagons and their degree of confinement at the tube ends.39,40 In accordance with theoretical predictions,41 CITS measurements on MWCNTs grown directly on the HOPG support for STM measurements42 showed that the MWCNTs with diameter over 30 nm do not have significant differences in their STS spectrum as compared to the HOPG support43 (Fig. 8).
Figure 7. Computed STM images: a) a single layer of graphene, which can be regarded as the equivalent of the SWCNT and b) HOPG, which can be regarded as the equivalent of MWCNT. (After Ref. 44).
STS was successfully used to investigate the effect of mechanical strain on the electronic structure of MWCNTs on a support. The spectroscopic data show that opposite to small deformations, high strain like that due to abrupt substrate morphology change showed drastic perturbation in the local electronic structure of the tube.45 This experimental finding is an agreement with theoretical work on the effects of defects produced by strain in the STM image of SWCNTs and alterations in their transport properties.46
29 4. STM and STS of singlewall carbon nanotubes The first atomic resolution topographic STM images and STS results on SWCNTs were reported simultaneously by two groups.47,48 Both groups used SWCNTs grown by laser ablation and the samples were imaged in a low temperature UHW STM on Au substrate. The atomic resolution images allow the determination of the chiral angle ), between the tube axis and the centers of the closest row of hexagons appearing in the STM (Fig. 2). As already discussed above, the hexagons will appear with a dark center (tunneling current minimum), while the bonds linking carbon atoms will show up as tunneling current maxima (Fig.7 and Fig. 9). However, caution is needed in the determination of the chiral angle from STM images, specially in the case of SWCNTs in bundles49 or nanotubes trapped in debris, like in Fig. 9, where the SWCNT may be twisted in a way that the angles between the tube axis T and the vector H (Fig. 2) may be altered with respect to those characteristic for a free tube. For a more detailed discussion see Ref. 49.
Figure 8. Topographic and CITS image of two multiwall carbon nanotubes on HOPG. The CITS image does not reveal differences in the electronic structure. In the STS plot, the IV curves recorded in the points marked by crosses are shown. In the range of –1 to 1 V all curves are practically coincident.
Another factor that may influence the precise determination of the chiral angle was identified by Venema et al.26 and is illustrated in Fig. 10. The distortion originates from the planar scan over a cylindrical object. Given the correct diameter and correct chirality, one may attempt the structural identification of the nanotube.2 On the basis of this identification, the
30 calculated and measured DOS may be compared. The electronic structure of the SWCNTs has one dimensional character, which means that in their electronic DOS are present sharp maxima, the socalled van Hove singularities.2 The STS results47,48 are in good agreement with the calculated densities of states of SWCNTs,2,50,51 the experimental DOS. The quantity dI/dV, called differential tunneling resistance, calculated according to Eq. 4, matches nicely the theoretical predictions. Both metallic and semiconductor carbon nanotubes were found. The typical forbidden gap values were in the range of 0.5 eV for semiconductor nanotubes with diameters in the 1.2 to 1.9 nm range, while the metallic carbon nanotubes with diameters in the range of 1.1 to 2.0 nm had subband separations of the order of 1.7 eV (Ref. [47]) (see also Fig. 11). The semiconducting SWCNTs were found to have gap values proportional to 1/dt, where dt is the diameter of the nanotube. In first approximation, the distances between van Hove singularities across the gap 'E11 of a semiconduting SWCNT are found to respect the rule 'E33 = 4 x 'E11 and 'E22 = 2 x 'E11, while for metallic nanotubes, the rule is 'E33 = 3 x 'E11 and 'E22 = 2 x 'E11 (Refs. [51,52]), 'E11 being the subband separation. This relationship may be very useful in the experimental identification SWCNTs from often noisy STS data. Due to work function differences, the Au substrate may produce the doping of the nanotube by charge transfer.47 Therefore, this effect has to be taken into account in the interpretation of the spectroscopic data. The center of the gap that separates the first two van Hove singularities may be offset towards the valence band by as much as 0.2 – 0.3 eV. Pronounced peaks in the LDOS were found at the end of an atomically resolved metallic SWCNT. Comparison of these data with calculations suggests that the topological arrangement of pentagons is responsible for the localized features in the experiment.53 Atomic resolution images of SWCNTs have been also obtained on the top of a rope.49,54,55 The achieved resolution made it possible to compare the chiralities of the tubes that belong to the same rope. In some cases it was reported that the nanotubes have different chiralities and some of them may be twisted. In other cases, nearly uniform diameter and chirality were observed. Spectroscopic data taken on bundles show that the IV curve may vary a lot along the same nanotube, and the spectra may exhibit a pronounced asymmetry.56 The variation of the noise level of constant current topographic images taken at the same values of positive and negative bias, confirm the asymmetric character of the spectra. The asymmetric shape of the STS spectra was explained by wavepacket dynamical simulations of the tunneling process23 showing that the asymmetry may arise from the shape of the electrode emitting the electrons.
31
Figure 9. Atomic resolution, constantcurrent STM images of a twisted zigzag nanotube trapped in surrounding material: a) the white line labeled AB is parallel with the tube axis while the line CD is passing through a row of hexagons (as seen in the line cut shown at the right hand side of the image), lines AB and CD are orthogonal, the black lines 11' and 22' are also drawn along rows of hexagons (dark spots) as clearly seen in the image but they make different angles with line CD, such a situation is indicating twist in the tube; b) physically zoomed detail of a), note a wellresolved hexagon close to the center of the image, the linecut along a CC bond is shown on right side, it indicates a slightly increased distance of 0.16 nm.
32
Figure 10. Illustration of the image distortion mechanism for nanotubes and a correction method. The 1nm bar indicates the scale for both images in a) and b): a) An uncorrected image with an angle between armchair and zigzag directions of 34° instead of 30°. This is attributed to a distortion effect, which stretches the atomic lattice in the direction perpendicular to the tube. The apparent angle I = 6°; b) The same image as a), corrected for the distortion. This is done by decreasing the image in the perpendicular direction to the tube axis to obtain a 30° difference between the zigzag and armchair directions. A chiral angle I = 5° is determined from this corrected image; c) a sketch illustrating the geometrical distortion mechanism. A nanotube with radius r lies in the xy plane with its axis in the x direction. When the STM tip tunnels to the side of the nanotube, an atom at position (x,y,z) is projected in the STM image at [x,(1+1h/r)y,z]. (After Ref. 26)
33
Figure 11. Structure and spectroscopy of a nanotube junction. (A) Atomicallyresolved STM image of a SWNT containing the junction; the position is highlighted with a white arrow. Black honeycomb meshes corresponding to (21,–2) and (22,–5) indices are overlaid on the upper and lower portions of the nanotube, respectively, to highlight the distinct atomic structures of these different regions. The image was recorded in a low temperature, UHV STM in constantcurrent mode with electrochemically etched tungsten tips at bias voltage Vb = 650 mV and It = 150 pA; (B) Tunneling conductance, dI/dV recorded at the upper (ź) and lower (Ÿ) locations indicated in (A). The energy difference between first VHS gap in the upper semiconducting segment, 0.45 eV, and lower metallic segment, 1.29 eV, are shown. (C) Spatially resolved dI/dV acquired across the junction at the positions indicated by the six symbols on the highresolution image (inset) of the junction interface. The small arrows highlight the positions of the first VHS of the semiconducting (21,22) structure and emphasize their spatial decay across the junction; the large arrows highlight the first VHS of the metallic (22,–5) structure. (From Ref. [67])
Figure 12. Constantcurrent (topographic) STM images calculated using a tightbinding Hamiltonian and a pointlike tip for an armchair (10,10), a chiral (13, 7), and a zigzag (18, 0) nanotube. (From Ref. [69])
34 The STM can be used as “tool” to modify the nanotube. SWCNTs were cut to desired size by applying a voltage pulse of the order of 4V, no polarity dependence was found.57 After shortening to a length of 30 nm, the STS spectrum of the nanotube piece taken at 4 K showed nonequidistant, steplike features absent before the cutting. The observed effect was interpreted as arising from discrete electron stationary waves observed as periodic oscillations in the differential conductance as a function of the position along the tube axis, with a period that differed from that of the atomic lattice. Similar studies were also devoted to chiral metallic and semiconducting nanotubes shortened to a few nanometers.58,59 In contrast to the metallic nanotubes, no significant length dependence is observed in finitesized semiconducting nanotubes down to 5 nm. The STS data obtained from the center of the shortened tubes showed a striking resemblance to the spectra observed before cutting.59 STS was used to investigate the effects of the ambient atmosphere on SWCNT samples. It was found that some semiconducting samples exhibit metallic behavior upon exposure to oxygen, while in the case of other samples the effect of O2 is restricted to the modification of the apparent density of states at the valence band edge.60 The investigation of the electrical response of semiconductor SWCNTs to gas molecules showed that while NH3 produces a sharp decrease in conductance, NO2 has an opposite effect.61 Statistical evaluation of nanotube diameters in a SWCNT mat (buckypaper) from the measured gap values and the comparison of these values with statistics obtained from TEM images demonstrated that the evaluation of diameters from gap values has to be done with caution because the gap is not always sharply defined in the STS spectrum.62 When examined in a large enough bias interval, the positions of the peaks in DOS corresponding to the van Hove singularities form a fingerprint of the wrapping indices n and m of the nanotube, which can be used to better estimate the diameter.63 Effects arising from interaction with the substrate64 and possible charge transfer47,65 have to be taken carefully into account. These effects may shift the Fermi level as mentioned above but do not affect too much the local DOS on the topmost part of the nanotube, which is the part that is scanned by the STM tip.64 Curvature effects in carbon nanotubes studied analytically as a function of chirality show that the S orbitals are found to be significantly rehybridized in all tubes so that they are never normal to the tube surface. This results in a curvatureinduced gap in the electronic band structure of the metallic tubes. The tilting of the S orbitals should be observable by atomic resolution scanning tunneling microscopy measurements.66 Recent STS measurements confirm the existence of the predicted gaps in the case of metallic zigzag nanotubes and the presence of pseudogaps produced by tubetube interaction in bundles.59,67
35 5. Computer modeling of STM images As already emphasized in the previous sections, the interpretation of experimental data obtained in a complex tunneling system, constituted of a supported nanoobject sandwiched between two tunneling gaps, is far from being straightforward. Therefore, computer modeling of the expected STM images is of great help in the correct evaluation of the experimental data. Modeling the atomic resolution images has been carried out by Meunier and Lambin. Their method is based on a tightbinding Selectron Hamiltonian for the case of a pointlike tip in combination with a selfsupported nanotube.68 The calculated images confirm the different image symmetries observed experimentally (see Fig. 12). In the case of armchair tubes, the honeycomb structure is easily observable with all bonds looking similar. For zigzag tubes, a strong anisotropy of the bonds is found with the ones parallel with the tube axis appearing much more stronger.26 For chiral tubes, stripes spiraling around the nanotube are observed19 which reverse orientation on bias reversal69,70 in the case of semiconducting tubes as indeed observed experimentally in some occasions.71 Calculated atomic resolution images of semiconductor SWCNTs show that the angle between the stripes spiraling around the nanotube and the tube axis changes gradually with the changing of the chiral angle. The two distinct classes of semiconductor tubes n – m = M(3) + 1 and n – m = M(3) – 1 exhibit complementary behavior. The image acquired with a negative tip in the first case is similar to the image acquired with positive tip bias for the second group of tubes. The complementarities of the two families of tubes is produced by the change in the occupied/unoccupied character of the HOMO and LUMO when changing from n – m = M(3) + 1 to n – m = M(3) –1 due to the phase change imposed by the periodic boundary condition along the circumferential direction.72 These calculated STM images of SWCNTs may prove to be extremely useful for the quick identification of nanotubes while acquiring experimental data. The method based on the tightbinding Selectron Hamiltonian is suitable for calculating the expected STM images of 5  7 defects and other defects on carbon nanotubes as well.68 An overview of the method and of the main results has been given recently.73 A similar calculation technique was used67 to identify the particular arrangement of the non hexagonal rings which may generate the intramolecular junction, seen in Fig. 11. A wavepacket dynamical method for the simulation of the geometric and charge spreading effects during tunneling in supported carbon nanostructures was developed and applied successfully for getting insight in the particularities of tunneling in twodimensional (2D) representations20 and more recently in full threedimensional (3D) tunneling through carbon nanotubes.74,75
36
Figure 13. Model geometries for the case of an infinite nanotube on flat substrate and for a capped nanotube protruding from a step. The time evolution clearly shows the different ways in which tunneling occurs. While at 2.54 fs after the wavepacket is launched, the two systems are are practically identical; at 8.47 fs major differences are produced, due to the existence of the cap and the absence of the substrate within tunneling range under the STM tip. (From Ref. [75])
In this method the current density flowing through the STM tipnanotubesupport system is calculated based on the scattering of the wavepackets incident on the barrier potential. The geometric effects are well described by electrodes and nanotube treated in the jelliumpotential approximation. The STM topographic profile through a carbon nanotube was calculated. It shows that as long as the electronic structure of the supported object and that of the support may be regarded as similar, the major image distortions arise from pure geometric tipshape convolution20 already discussed above. With increasing differences in the electronic structure of the nanotube and that of the support, larger distortions are expected. The incidence angle and energydependent transmission of wavepackets through single and bundled carbon nanotubes were calculated, the effects arising from pointcontact imaging and tip shape were analyzed, and STS curves were modeled.23 As already mentioned above, for tip positive bias, the angular dependence of the transmission depends strongly on the nature of the nanosystem in the STM gap. While the transmission of the STM tunnel junction without nanotube can be well represented by a one dimensional model, all other geometries cause large normaltransverse momentum mixing of the wave packet. Recently, an overview and comparison of the tightbinding and wave packet dynamical methods has been given.36 The usefulness of the wavepacket dynamical method is completely revealed when it is used in full threedimensional (3D) simulation.36,74,75
37
Figure 14. Topographic STM images of defects created on a MWCNT by irradiation with 30 keV Ar+ ions: a) defects created by irradiation so called "hillocks"; b) linecut along the line in a) on may note that an apparent height of around 1 nm is measured for the defects. (From Ref. [79])
The computer simulation clearly shows that during tunneling, the charge spreads along the nanotube. As a result, while the tunnel junction between the STM tip and the nanotube is a zerodimensional junction, the tunnel junction between the nanotube and the support will be a linear onedimensional junction (Fig. 13). The consequences of the spreading will affect the tunneling situations when the nanotube is only partly supported or is placed over a nonhomogeneous substrate.74,75 6.
Structural defects
As already reveled by early theoretic papers, structural defects like 5 –7 pairs, vacancies, isolated pentagons and heptagons may have dramatic effects on the structure and electronic properties of carbon nanotubes.58,76,77 More recent calculations78 predict similar effects for defects produced by ion irradiation. Irrespective of its nature, a defect in a SWCNT gives rise to backscattering of the electron wave function. Interference of the incoming and backscattered waves leads to a standing wave pattern, which can extend far away from the defect because of the onedimensionality of the nanotube. This is the case, for example, for the junction of the two tubes, shown in Fig. 11 or for defects created on purpose by ion irradiation79 (Fig. 14).
38 7. Summary and outlook Scanning tunneling microscopy and spectroscopy proved to be the most suitable methods for truly revealing the characteristic electronic properties of carbon nanotubes arising from their onedimensional character and nanometric size. A remarkable agreement between the theoretically computed results and experiment was achieved in the case of SWCNTs. Many basic level questions concerning the STM investigation of multiwall carbon nanotubes are still open. The advance of this field is expected as suitable techniques emerge for the selective production of few wall carbon nanotubes like double wall carbon nanotubes. The variety of novel carbon nanostructures, not discussed in this chapter like Yjunctions, coils, and multiple coils, will pose a new challenge both for theoreticians and experimentalists.8082 A new class of carbon nanomaterials emerges with the discovery of the peapods. The first experimental results already indicate that STM will be an extremely useful tool in getting deeper insight in the properties of these new materials.83 Last but not least, a very wide field of research opens up with the functionalization of carbon nanotubes. As the first results already demonstrate, STM and STS are the ideal tools for understanding the way in which the chemical moieties are coupled to the nanotubes.84,85 A close collaboration between physicists, chemists and materials scientist will be needed in this field. ACKNOWLEDGMENT
This work was partly funded by the InterUniversity Attraction Pole (Grant No. IUAP P5/1) on "Quantumsize effects in nanostructured materials" of the Belgian Science Programming Services and partly by OTKA Grant T 043685 in Hungary. L.P.B. gratefully acknowledges the Belgian Fonds National de la Recherche Scientifique and the Hungarian Academy of Sciences for financial support.
References 1. 2. 3.
H. W. Kroto, J. R. Heath, S. C. O’Brien, R. F. Curl, and R. E. Smalley, C60: Buckminsterfullerene, Nature 318, 162163 (1985). M. S. Dresselhaus, G. Dresselhaus, and P. C. Eklund, Science of Fullerenes and Carbon Nanotubes, (Academic Press, San Diego, 1996). Carbon Filaments and Nanotubes: Common Origins, Different Applications? edited by L. P. Biró, C. A. Bernardo, G. G. Tibbetts and Ph. Lambin (Kluwer Acad. Publ., Dordrecht, 2001).
39 4. 5. 6. 7. 8. 9.
10. 11.
12.
13.
14. 15. 16. 17. 18. 19. 20.
21. 22.
23.
24. 25.
S. Iijima, Helical microtubules of graphitic carbon, Nature 354, 56 58 (1991). G. E. Scuseria, Negative curvature and hyperfullerenes, Chem. Phys. Lett. 195, 534536 (1992). L. A. Chernozatonskii, Carbon nanotube connectors and planar jungle gyms, Phys. Lett. A 172, 173 (1992). B. I. Dunlap, Connecting carbon tubules, Phys. Rev. B 46, 19331936 (1992). S. Ihara, S. Itoh, and J. Kitakami, Helically coiled cage forms of graphitic carbon, Phys. Rev. B 48, 56435647 (1993). S. Amelinckx, X. B. Zhang, D. Bernaerts, X. F. Zhang, V. Ivanov, and J. B. Nagy, A formation mechanism for catalytically grown shaped graphite nanotubes, Science 265, 635639 (1994). J. Li, C. Papadopoulos, and J. M. Xu, Nanoelectronics: Growing Yjunction carbon nanotubes, Nature 402, 253254 (1999). L. P. Biró, S. D. Lazarescu, P. A. Thiry, A. Fonseca, J. B. Nagy, A. A. Lucas, and Ph. Lambin, Scanning tunneling microscopy observation of tightly wound, singlewall coiled carbon nanotubes, Europhys. Lett. 50, 494 (2000). L. P. Biró, R. Ehlich, Z. Osváth, A. Koós, Z. E. Horváth, J. Gyulai, and J. B.Nagy, Room temperature growth of singlewall coiled carbon nanotubes and Ybranches, Mat. Sci. Eng. C 19, 3 (2002). L. P. Biró, G. I. Márk, A. A. Koós, J. B.Nagy, Ph. And Lambin, Coiled carbon nanotube structures with supraunitary nonhexagonal to hexagonal ring ratio, Phys. Rev. B 66, 165405 (2002). D. Reznik, C. H. Olk, D. A. Neumann, and J. R. D. Copley, Xray powder diffraction from carbon nanotubes and nanoparticles, Phys. Rev. B 52, 116124 (1995). Ph. G. Collins, M. S. Arnold, and Ph. Avouris, Engineering Carbon Nanotubes and Nanotube Circuits Using Electrical Breakdown, Science 292, 706709 (2001). R. Wiesendanger, Scanning Probe Microscopy and Spectroscopy, Cambridge University Press, Cambridge (1994). J. Tersoff and D. R. Hamann, Theory of the scanning tunneling microscope, Phys. Rev. B 31, 805813 (1985). R. Wiesendanger and H.J. Güntherodt (Editors), Scanning Tunneling Microscopy III, SprigerVerlag, (1993). W. Clauss, Scanning tunneling microscopy of carbon nanotubes, Appl. Phys. A 69, 275281 (1999). G. I. Márk, L. P. Biró, and J. Gyulai, Simulation of STM images of threedimensional surfaces and comparison with experimental data: Carbon nanotubes, Phys. Rev. B 58, 1264512648 (1998). Animated computer simulations showing the tunneling through a supported nanoobject are available at: http://www.mfa.kfki.hu/int/nano/ L. P. Biró, J. Gyulai, Ph. Lambin, J. B.Nagy, S. Lazarescu, G. I. Márk, A. Fonseca, P. R. Surján, Zs. Szekeres, P. A. Thiry, and A. A. Lucas, Scanning tunneling microscopy (STM) imaging of carbon nanotubes, Carbon 36, 689696 (1998). G. I. Márk, L. P. Biró, J. Gyulai, P. A. Thiry, A. A. Lucas, and Ph. Lambin, Simulation of scanning tunneling spectroscopy of supported carbon nanotubes, Phys. Rev. B 62, 27972805 (2000). L. P. Biró and G. Márk, STM investigation of Carbon Nanotubes, in Ref. 3, pp. 219231 L. P. Biró, S. Lazarescu, Ph. Lambin, P. A. Thiry, A. Fonseca, J. B. Nagy, and A. A. Lucas, Scanning tunneling microscope investigation of carbon nanotubes produced by catalytic decomposition of acetylene, Phys. Rev. B 56, 12 49012498 (1977).
40 26. L. C. Venema, V. Meunier, Ph. Lambin, and C. Dekker, Atomic structure of carbon nanotubes from scanning tunneling microscopy, Phys. Rev. B 61, 29912996 (2000). 27. Ph. Kim, T. W. Odom, J. Huang, and Ch. M. Lieber, STM study of singlewalled carbon nanotubes, Carbon 38, 17411744 (2000). 28. F.X. Zha, D. L. Carroll, R. Czerw, A. Loiseau, H. Pascard, W. Clauss, and S. Roth, Electronic effects in scanning tunneling microscopy of dendritic, Crfilled carbon nanotubes, Phys. Rev. B 63, 165432165437 (2001). 29. Z. Zang and Ch. M. Lieber, Nanotube structure and electronic properties probed by scanning tunneling microscopy, Appl. Phys. Lett. 62, 27922794 (1993). 30. Ch. H. Olk and J. P. Heremans, Scanning tunneling spectroscopy of carbon nanotubes, J. Mater. Res. 9, 259262 (1994). 31. J. W. Mintmire, B. I. Dunlap, and C. T. White, Are fullerene tubules metallic? Phys. Rev. Lett. 68, 631634 (1992). 32. R. Saito, M. Fujita, G. Dresselhaus, and M. S. Dresselhaus, Electronic structure of chiral graphene tubules, Appl. Phys. Lett. 60, 22042206 (1992). 33. Y. K. Kwon and D. Tománek, Electronic and structural properties of multiwall carbon nanotubes, Phys. Rev. B 58, R16001R16004 (1998). 34. M. Ge and K. Sattler, Vaporcondensation generation and STM analysis of fullerene tubes, Science 260, 515518 (1993). 35. D. Tománek and S. G. Louie, Firstprinciples calculation of highly asymmetric structure in scanningtunnelingmicroscopy images of graphite, Phys. Rev. B 37, 83278336 (1988). 36. G. I. Márk, L. P. Biró, and Ph. Lambin, in Frontiers of Multifunctional Nanosystems, edited by E. Buzaneva and P. Scharff, Kluwer Academic Publishers, Dordrecht (2002), p. 43. 37. J. Jxhie, K. Sattler, M. Ge, N. Verkateswaran, Giant and supergiant lattices on graphite, Phys Rev. B 47, 1583515841 (1993). 38. Ph. Lanbin, V. Meunier, and A. Rubio, Simulation of STM images and STS spectra of carbon nanotubes, in: "Science and application of nanotubes" Ed. D. Tomanek and R. J. Endbody (Kluwer Academic/Plenum Publishers, New York, 2000) pp. 17 33. 39. A. Hassanien, A. Mrzel, M. Tokumoto, and D. Tománek, Imaging the interlayer interactions of multiwall carbon nanotubes using scanning tunneling microscopy and spectroscopy, Appl. Phys. Lett. 79, 42104212 (2001). 40. D. L. Carroll, P. Redlich, P. M. Ajayan, J. C. Charlier, X. Blase, A. De Vita, and R. Car, Electronic Structure and Localized States at Carbon Nanotube Tips, Phys. Rev. Lett. 78, 28112814 (1997) 41. T. Tamura and M. Tsukada, Electronic states of the cap structure in the carbon nanotube, Phys. Rev. B 52, 60156026 (1995). 42. M. S. Dresselhaus, in Carbon Filaments and Nanotubes: Common Origins, Different Applications? edited by L. P. Biró, C. A. Bernardo, G. G. Tibbetts and Ph. Lambin, Kluwer Academic Publishers, Dordrecht (2001), p. 11 43. L. P. Biró, R. Ehlich, R. Tellgmann, A. Gromov, N. Krawez, M. Tschaplyguine, M.M. Pohl, Z. Véretsy, Z. E. Horváth, and E. E. B. Campbell, (1999) Growth of carbon nanotubes by fullerene decomposition in the presence of transition metals, Chem. Phys. Lett. 306, 155 – 162. 44. L. P. Biró and G. I. Márk in Carbon Filaments and Nanotubes: Common Origins, Different Applications? edited by L. P. Biró, C. A. Bernardo, G. G. Tibbetts and Ph. Lambin, Kluwer Academic Publishers, Dordrecht (2001), p. 219 45. D. Tekleab, D. L. Carroll, G. G. Samsonidze, and B. I. Yakobson, Straininduced electronic property heterogeneity of a carbon nanotube, Phys. Rev. B 64, 035419035424 (2001).
41 46. D. Orlikowski, M. B. Nardelli, J. Bernholc, and Ch. Roland, Theoretical STM signatures and transport properties of native defects in carbon nanotubes, Phys. Rev. B 61, 1419414203 (2000). 47. J. W. G. Wildöer, L. C. Venema, A. G. Rinzler, R. E. Smalley, and C. Dekker, Electronic structure of atomically resolved carbon nanotubes, Nature 391, 5962(1998) 48. T. W. Odom, JL. Huang, Ph. Kim, and Ch. M. Lieber, Atomic structure and electronic properties of singlewalled carbon nanotubes, Nature, 391, 6264 (1998). 49. W. Clauss, D. J. Bergeron, and A. T. Johnson, Atomic resolution STM imaging of a twisted singlewall carbon nanotube, Phys. Rev. B 85, R4266R4269 (1998). 50. N. Hamada, S. Sawada, and A. Oshiyama, New onedimensional conductors: Graphitic microtubules, Phys. Rev. Lett. 68, 15791581 (1992). 51. J.C. Charlier and Ph. Lambin, Electronic structure of carbon nanotubes with chiral symmetry, Phys. Rev. B 57, R15037R15039 (1998). 52. C. T. White and J. W. Mintmire, Density of states reflects diameter in nanotubes, Nature 394 2930 (1998). 53. Ph. Kim, T. W. Odom, JL. Huang, and Ch. M. Lieber, Electronic Density of States of Atomically Resolved SingleWalled Carbon Nanotubes: Van Hove Singularities and End States, Phys. Rev. Lett. 82, 12251228 (1999). 54. E. D. Obraztsova, V. Yu. Yurov, V. M. Shevluga, R. E. Baranovsky, V. A. Nalimova, V. L. Kuznetsov, and V. I. Zaikovski, Structural investigations of closepacked singlewall carbon nanotube material, Nanostruct. Mater. 11, 295306 (1999). 55. A. Hassanien, M. Tokumoto, Y. Kumazawa. H. Kataura, Y. Maniwa, S. Suzuki, and Y. Achiba, Atomic structure and electronic properties of singlewall carbon nanotubes probed by scanning tunneling microscope at room temperature, Appl. Phys. Lett. 73, 38393841 (1998). 56. L. P. Biró, P. A. Thiry, Ph. Lambin, C. Journet, P. Bernier, and A. A. Lucas, Influence of tunneling voltage on the imaging of carbon nanotube rafts by scanning tunneling microscopy, Appl. Phys. Lett. 73, 36803682 (1998). 57. L. C. Venema, J. W. G. Wildöer, H. L. J. Temminck Tuinstra, C. Dekker, A. G. Rinzler, and R. E. Smalley, Length control of individual carbon nanotubes by nanostructuring with a scanning tunneling microscope, Appl. Phys. Lett. 71, 26292631 (1977). 58. T. W. Odom, J. L. Huang, Ph. Kim, and C. M. Lieber, Structure and Electronic Properties of Carbon Nanotubes, J. Phys. Chem. B, 104, 27942809 (2000). 59. T. W. Odom, J. L. Huang, and C. M. Lieber, STM studies of singlewalled carbon nanotubes, J. Phys. Condens. Mat. 14, R145R167 (2002). 60. Ph. G. Collins, K. Bradley, M. Ishigami, and A. Zettl, Extreme Oxygen Sensitivity of Electronic Properties of Carbon Nanotubes, Science 287, 18011804 (2000). 61. J. Kong, N. R. Franklin, Ch. Zhou, M. G. Chapline, S. Peng, K. Cho, and H. Dai, Nanotube Molecular Wires as Chemical Sensors, Science 287, 622625 (2000). 62. I. Wirth, S. Eisebitt, G. Kann, and W. Eberhardt, Statistical analysis of the electronic structure of singlewall carbon nanotubes, Phys. Rev. B 61, 57195723 (2000). 63. R. Saito, G. Dresselhaus, and M. S. Dresselhaus, Trigonal warping effect of carbon nanotubes, Phys. Rev. B 61, 29812990 (2000). 64. A. Rubio, Spectroscopic properties and STM images of carbon nanotubes, Appl. Phys. A 68, 275282 (1999). 65. Y. Xue and S. Datta, FermiLevel Alignment at MetalCarbon Nanotube Interfaces: Application to Scanning Tunneling Spectroscopy, Phys. Rev. Lett. 83, 48444847 (1999). 66. A. Kleiner and S. Eggert, Curvature, hybridization, and STM images of carbon nanotubes, Phys. Rev. B, 64, 113402113406 (2001).
42 67. M. Ouyang, JL. Huang, C. L. Cheung, and Ch. M. Lieber, Energy Gaps in "Metallic" SingleWalled Carbon Nanotubes, Science 292, 702705 (2001). 68. V. Meunier and Ph. Lambin, TightBinding Computation of the STM Image of Carbon Nanotubes, Phys. Rev. Lett. 81, 55885591 (1999). 69. Ph. Lambin, A. Loiseau, C. Culot, and L. P. Biró, Structure of carbon nanotubes probed by local and global probes, Carbon 40, 16351648 (2002). 70. C. L. Kane and E. J. Mele, Broken symmetries in scanning tunneling images of carbon nanotubes, Phys. Rev. B 59, R12759R12762 (1999). 71. W. Clauss, D. J. Bergeron, M. Freitag, C. L. Kane, E. J. Mele, and A. T. Johnson, Electron backscattering on singlewall carbon nanotubes observed by scanning tunneling microscopy, Europhys. Lett. 47, 601607 (1999). 72. P. Lambin, G. I. Márk, V. Meunier, L. P. Biró, Computation of STM Images of Carbon Nanotubes, Int. J. Quant. Chem. 95, 493–503 (2003). 73. Ph. Lambin and V. Meunier, in Carbon Filaments and Nanotubes: Common Origins, Different Applications? edited by L. P. Biró, C. A. Bernardo, G. G. Tibbetts and Ph. Lambin, Kluwer Academic Publishers, Dordrecht (2001), p. 233. 74. G. I. Márk , A. Koós , Z. Osváth , L. P. Biró , J. Gyulai , A. M. Benito , W. K. Maser, P. A. Thiry, and Ph. Lambin, Calculation of the charge spreading along a carbon nanotube seen in scanning tunnelling microscopy (STM), Diam. Rel. Mat. 11, 961963 (2001). 75. G. I. Márk, Ph. Lambin, and L. P. Biró, Calculation of axial charge spreading in carbon nanotubes and nanotube Y junctions during STM measurement, Phys. Rev B 70, 115423/110 (2004). 76. A. L. Macky and H. Terrones, Diamond from graphite, Nature 352, 762762 (1991). 77. 77 JC. Charlier, T. W. Ebbesen, and Ph. Lambin, Structural and electronic properties of pentagonheptagon pair defects in carbon nanotubes, Phys. Rev B 53, 1110811113 (1996). 78. A. V. Krasheninnikov, K. Nordlund, M. Sirviö, E. Salonen, and J. Keinonen, Formation of ionirradiationinduced atomicscale defects on walls of carbon nanotubes, Phys. Rev. B 63, 245405245411 (2001). 79. Z. Osváth, G. Vértesy, L. Tapasztó, F. Wéber, Z. E. Horváth, J. Gyulai, and L. P. Biró, Atomically resolved STM images of carbon nanotube defects produced by Ar irradiation, Phys. Rev B 72, 045429/16 (2005). 80. Z. Osváth, A. A. Koós, Z. E. Horváth, J. Gyulai, A. M. Benito, M.T. Martínez, W. K. Maser, and L. P. Biró, Arcgrown Ybranched carbon nanotubes observed by scanning tunneling microscopy (STM), Chem. Phys. Lett. 365, 338342 (2002). 81. L. P. Biró , R. Ehlich , Z. Osváth , A. Koós , Z. E. Horváth , J. Gyulai , and J. B. Nagy, From straight carbon nanotubes to Ybranched and coiled carbon nanotubes, Diam. Rel. Mat. 11, 10811085 (2002). 82. Ph. Lambin, G. I. Márk and L. P. Biró, Structural and electronic properties of coiled and curled carbon nanotubes having a large number of pentagon–heptagon pairs, Phys. Rev B 67, 205413/19 (2003). 83. D. J. Hornbaker, S.J. Kahng, S. Misra, B. W. Smith, A. T. Johnson, E. J. Mele, D. E. Luzzi, and A. Yazdani, Mapping the OneDimensional Electronic States of Nanotube Peapod Structures, Science 295, 828831(2002). 84. K. F. Kelly, I. W. Chiang, E. T. Mickelson, R. H. Hauge, J. L. Margrave, X. Wang, G. E. Scuseria, C. Radloff, N. J. Halas, Insight into the mechanism of sidewall functionalization of singlewalled nanotubes: an STM study, Chem. Phys. Lett. 313, 445450 (1999). 85. Z. Kónya, I. Vesselenyi, K. Niesz, A. Demortier, A. Fonseca, J. Delhalle, Z. Mekhalif, J. B.Nagy, A. A. Koós, Z. Osváth, A. Kocsonya, L. P. Biró, and I. Kiricsi, Large scale production of short functionalized carbon nanotubes, Chem. Phys. Lett. 360, 429435(2002).
STRUCTURAL DETERMINATION OF INDIVIDUAL SINGLEWALL CARBON NANOTUBE BY NANOAREA ELECTRON DIFFRACTION E. THUNE, D. PREUSCHE, C. STRUNK* Institut für Experimentelle und Angewandte Physik, Universität Regensburg, Universitätsstr. 31, D93040 Regensburg, Germany H. T. MAN, A. MORPURGO Kavli Institute of NanoScience, Delft University of Technology, Lorentzweg 1, NL2628 CJ Delft, The Netherlands F. PAILLOUX Laboratoire de Métallurgie Physique, Université de Poitiers, Bd M. & P. Curie, F86962 ChasseneuilFuturoscope Cedex, France A. LOISEAU Laboratoire d’Etudes des Microstructures, CNRSONERA, 29 avenue de la Division Leclerc, F92322 Châtillon Cedex, France
Abstract. The determination of the exact values of (n,m) or equivalently (d, ) for an individual singlewall nanotube (SWNT) is of great importance, in particular for comparison with transport measurements. In this study, the selected area electron diffraction technique is used to investigate the structure of individual carbon nanotubes grown by chemical vapor deposition (CVD) directly on TEMSi3N4 membranes. Keywords: SWNT; SAED; chiral indices
As grown, singlewall nanotubes (SWNTs) usually have a dispersion of chirality and diameter1 because the nanotube structural energy is only weakly dependent on chirality.2 For a quantitative comparison of the measured electric transport properties with theoretical predictions, it is highly desirable to identify the chiral indices of individual SWNTs. A complete determination of the structure (diameter and helicity) with enough spatial resolution to identify an individual *To whom correspondence should be addressed. Christoph Strunk; email: [email protected]
43 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 43–44. © 2006 Springer. Printed in the Netherlands.
44 nanotube is in practice a difficult task. The indices (n,m) of an individual nanotube can be identified using selected area electron diffraction (SAED).3 In this study, we grow carbon SWNTs directly on TEMSi3N4 membranes previously decorated by electron beam lithography (EBL) with (a) a pattern of gold (Au) alignment marks in order to locate individual nanotubes, (b) with a pattern for catalyst particles for nanotube growth by chemical vapor deposition (CVD), and (c) with a pattern of ironpalladium (FePd) contact structures. Atomic Force Microscopy (AFM) (Fig. 1, left) and High Resolution Transmission Microscopy (HRTEM) (Fig. 1, middle) characterizations were achieved and showed the presence of isolated SWNTs, 15 nm in diameter, with low contamination of amorphous carbon. The background from the amorphous Si3N4 membrane covers the weak signal of individual SWNT (see Fig. 1, middle) and does not allow us to determine the chirality of the nanotubes.
Figure 1. Left: AFM height image of isolated nanotubes on Si3N4 membrane. Middle: HRTEM image of an individual SWNT grown on Si3N4 membrane and its corresponding diffraction pattern. Right: TEM image of an individual SWNT on a perforated Si3N4 membrane and its corresponding diffraction pattern.
To avoid the Si3N4 background, the selected area electron diffraction requires an energy filter or use perforated membranes. Figure 1, right, shows an individual SWNT (5.5 nm in diameter) grown on perforated membranes. But, up to now, since individual SWNTs produce very low intensities, it was difficult to determine their chiral indices. References 1. 2. 3.
R. Saito, G. Dresselhaus, and M. S. Dresselhaus, Physical properties of carbon nanotubes (Imperial College Press, London, 1998). J. W. Mintmire and C. T. White, Electronic and structural properties of carbon nanotubes, Carbon 33, 893 902 (1995). M. Gao, M. Zuo, R. D. Twesten, I. Petrov, L. A. Nagahara, and R. Zhang, Structure determination of individual singlewall carbon nanotubes by nanoarea electron diffraction, Appl. Phys. Lett. 82, 27032705 (2003).
THE STRUCTURAL EFFECTS ON MULTIWALLED CARBON NANOTUBES BY THERMAL ANNEALING UNDER VACUUM K. D. BEHLER, H. YE, S. DIMOVSKI, Y. GOGOTSI Drexel University, Philadelphia, PA 19103, USA
Abstract. Characterization of Vacuum Thermal Annealing of a new variety of Multiwalled Carbon Nanotubes (MWCNT) produced in a Catalytic Chemical Vapor Deposition (CCVD) process by Arkema, France, was preformed by Raman Spectroscopy (RS) and Transmission Electron Microscopy (TEM). Keywords: Nanotube; Raman; Transmission Electron Microscopy; Annealing
Recently, small tubes, having advantages with respect to mechanical and thermal properties over larger tubes, expect to see manufacturing capabilities allowing large scale quantities that may dominate the market within 23 years. The effects of heat treatment were noticeable for the nanotubes when compared to the asreceived tubes. The tubes were vacuum annealed (106 atm) for 3 hours at 1800°C and 2000°C. Catalyst removal can be seen in Figure 1. The asreceived tubes encompass the metal catalyst (Fig. 1a); whereas catalyst removal occurs after vacuum annealing (Fig. 1b).
Figure 1. TEM of MWCNTs.
Individual tubes appear to be more graphitic after the 1800°C treatment and even more so after the treatment at 2000°C as evidenced by the RS, Figure 2 45 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 45–46. © 2006 Springer. Printed in the Netherlands.
46 (excitation energy of 1.96 eV). A separation of the G and D' begins to occur, as seen in the inset, as the treatment temperature increases, implying that the tubes are transforming to a more graphitic structure.
Figure 2. Raman Spectra of MWCNTs.
Some tubes after treatment also show overgrowth, as shown in Figure 1c. The original tube, which is currently in the center of tube, has a sheathlike growth that is enclosing the original structure. Thickening of the tubes indicates mass transfer occurs during the annealing process. The observed tubes show a polygonal graphite structure as the tube walls start to become less cylindrical (Fig 1d) similar to graphite polyhedral crystals.1 There is also an elimination of dangling bonds nanotube tips through the “lip+lip” interactions as seen in the inset of Figure 1c. The graphene sheets are connecting to other sheets and forming a loop,2 preventing bonding at the end of the tube by reducing the number of possible bonding sites and thus eliminating dangling bonds. Such nanotubes are expected to be less chemically active. Vacuum annealing leads to catalyst removal, polygonization and graphitization of MWCNT. It may lead to higher conductivity and improved electrical properties. Excessive over growth and polygonization occur above 1800°C. References 1. 2.
Y. Gogotsi, J. A. Libera, N. Kalashnikov, and M. Yoshimura, Graphite polyhedral crystals. Science 290, 317320 (2000). S. Rotkin and Y. Gogotsi, Analysis of nonplanar graphitic structures: from arched edge planes of graphite to nanotubes, Mat. Res. Innovat. 5, 191200 (2002).
TEM SAMPLE PREPARATION FOR STUDYING THE INTERFACE CNTSCATALYSTSUBSTRATE MARIEFAITH FIAWOO, ANNICK LOISEAU LEM UMR 104 CNRS/ONERA, 92322 Châtillon, France ANNEMARIE BONNOT, ANTONIO IAIA LEPESCNRS UPR11, 38042 Grenoble, France VINCENT BOUCHIAT CRTBTCNRS UPR5001, 38042 Grenoble, France JANY THIBAULT TECSEN CNRS UMR 6122, Univ. Paul Cézanne, 13397 Marseille, France
Abstract. We study by Transmission Electron Microscopy (TEM) insitu selfassembled carbon nanotubes (CNTs) grown by the Hot Filament Chemical Vapor Deposition (HFCVD) process on silicon substrates used for electronic devices. We present the different methods we have developed for extracting CNTs and studying their interface with the substrate as they are observable by TEM.
Keywords: TEM preparation technique, nanotubes, SACT, cleavage, ion milling
For studying specimens by TEM and performing analysis (EFTEM, EELS), the sample needs to be thin (20 to 30 nm), transparent to electrons, smaller than 3 nm diameter to fit in the TEM holder and keep its intrinsic features. We want to study the interface CNTscatalystsubstrate in localized CNTs synthesis (Marty et al., 2002) aiming to understand the CNTs growth by Chemical Vapor Deposition (CVD) process like previously done for high temperature synthesis (Loiseau et al., 2004). Here we present different preparation methods which we tried in order to reach the required thickness necessary for the study of the interface of nanotubes with a single crystal silicon substrate. Four types of preparation have been tested: scratching the sample, ion milling, Si3N4 membrane system, and small angle cleavage technique (SACT). *To whom correspondence should be addressed. MarieFaith Fiawoo; email: [email protected]
47 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 47–48. © 2006 Springer. Printed in the Netherlands.
48 Some of them enable plane view, crosssection observations or both. Scratching the sample with a tweezer or a diamond scribe to deposit on a TEM grid seldom enables observations of the interface because of the thickness of the piece scratched. Ion milling technique involves a mechanical polishing before thinning the sample by pulling out atoms of the sample with an ion beam of argon Ar+ (Fig. 1, left). A disadvantage of this technique is the introduction of artifacts. The Si3N4 membrane with a rectangular hole has CNTs grown on the membrane as well as across the hollow. Unfortunately, the hole's edge in not thin enough to see the interface. SACT consists in cleaving the sample along the (120) and (110) planes, which form an angle of 18.43° (Fig. 1, right). This small angle of the apex of the sample allows its transparency to electrons (Walck et al., 1997). The disadvantage of this technique is the small area transparent to electrons, which is not adapted for a sample having a few CNTs.
Figure 1. Left: Ion milling preparation stages; Right: Different stages of SACT.
For the time being, the ion milling technique is the only one, which enables the systematical observation and analysis of the interface. Further investigations, such as focused ion beam cutting and other membrane systems, are presently being carried out. References Marty, L., Bouchiat, V., Bonnot, A.M., Chaumont, M., Fournier, T., Decossas, S., and Roche, S. 2002, Batch processing of nanometerscale electrical circuitry based on insitu grown singlewalled carbon nanotubes, Microelectronic Engineering 6162:485489. Loiseau, A., Gavillet, J., Ducastelle, F., Thibault, J., Stéphan, O., Bernier, P., and Thair, S., 2004, Nucleation and growth of SWNT: TEM studies of the role of the catalyst, C. R. Physique 4:975991. Walck, S.D., and McCaffrey, J.P, 1997, The small angle cleavage technique applied to coatings and thin films, Thin Solid Films 308309:399405.
A METHOD TO SYNTHESIZE AND TAILOR CARBON NANOTUBES BY ELECTRON IRRADIATION IN THE TEM R. CAUDILLO,* M. JOSÉYACAMAN Materials Science and Engineering and Texas Materials Institute, University of Texas at Austin, 1 University Station C2201, Austin, Texas 78712, USA H. E. TROIANI Centro Atómico Bariloche and Instituto Balseiro, CNEA and UNC, 8400 Bariloche, Rio Negro, Argentina M. A. L. MARQUES, A. RUBIO Donostia International Physics Center, Paseo Manuel Lardizábal 4, 20018 San Sebastián, Spain
Abstract. Carbon nanotubes (CNTs) can be synthesized from amorphous carbon thin films by electron irradiation in the transmission electron microscope (TEM). We propose that irradiation effects impart a tensile stress and introduce defects to a CNT thus formed and show that the CNTs can be made to fracture by a brittle or ductile mechanism. We also report a healing of the carbon film that is observed after some critical amount of irradiation.
TEM is normally only used to characterize CNTs synthesized by standard techniques, such as arc discharge, laser ablation, or chemical vapor deposition (CVD); however, Troiani et al.1,2 have shown that it is also possible to use electron irradiation in a TEM to construct singlewalled carbon nanotubes (SWCNT) from an amorphous carbon film. In this method, two throughholes are created in an amorphous carbon film using electronbeam irradiation in a TEM. A thin graphitic fiber forms in between the two holes and is narrowed until forming a SWCNT. Here we show that by controlling the current density on a CNT formed by this technique, it is possible to cause fracture and to observe the fracture mechanism in the TEM. We observe that the CNT fractures either by a brittle or ductile fracture mechanism, depending on the amount of current density on the sample. *To whom correspondence should be addressed. Roman Caudillo; email: [email protected]
49 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 49–50. © 2006 Springer. Printed in the Netherlands.
50 We propose two mechanisms working to achieve the narrowing of the bridge: (1) deformation in response to an axial stress resulting from a contraction of the film due to radiationinduced graphitization and (2) creation of defects in the bridge due to irradiation, resulting in vacancy clustering that leads to brittle fracture or atomic rearrangements that lead to plastic fracture. Figure 1 shows CNT formation at the neck of a bridge and the formation of a linear chain of carbon atoms during the ductile deformation of the CNT.
Figure 1. (a) TEM image showing the formation of a CNT at the neck of a bridge, (b) TEM image showing the formation of a linear chain of carbon atoms (signaled by arrow) in a CNT neck that has undergone ductile deformation.
We also observe a surprising phenomenon during prolonged irradiation of carbon nanotubes and carbon nanofibers formed by electron irradiation in the TEM, in which a sudden diffusion of carbon atoms to the bridge structure and periphery of the holes results in a reconstruction of the carbon film, which in some cases results in a complete closing of the holes.3 These results indicate that CNTs, as well as other carbon nanostructures, can be synthesized and tailored by electron irradiation, and suggests the possibility of larger scale, carbon nanostructure design by ebeam lithography and complementary techniques. References 1. H. E. Troiani, M. MikiYoshida, G. A. CamachoBragado, M. A. L. Marques, A. Rubio, J. A. Ascencio, and M. JoseYacaman, Direct observation of the mechanical properties of singlewalled carbon nanotubes and their junctions at the atomic level, Nano Letters 3(6), 751755 (2003). 2. M. A. L. Marques, H. E. Troiani, M. MikiYoshida, M. JoseYacaman, and A. Rubio, On the breaking of carbon nanotubes under tension, Nano Letters 4(5), 811815 (2004). 3. R. Caudillo, H. E. Troiani, M. MikiYoshida, M. A. L. Marques, A. Rubio, and M. JoseYacaman, A viable way to tailor carbon nanomaterials by irradiationinduced transformations, Radiation Physics and Chemistry (In Press) (2005).
SCANNING TUNNELING MICROSCOPY STUDIES OF NANOTUBELIKE STRUCTURES ON THE HOPG SURFACE
I. N. KHOLMANOV,1,2 M. FANETTI,1 L. GAVIOLI,1 1 1,3 M. CASELLA, M. SANCROTTI 1
INFM and Dipartimento di Matematica e Fisica, Università Cattolica del Sacro Cuore, Via dei Musei 41, 25121 Brescia, Italy 2 Thermophysics Department, Academy of Sciences, Katartal str., 28, 700135 Tashkent, Uzbekistan 3 Laboratorio Nazionale TASCINFM, Strada Statale 14, Km 163.5, Basovizza, I34012 Trieste, Italy
Abstract. We have studied the scanning tunneling microscopy tip interaction with the naturally formed nanotubelike (NTL) structures on highly oriented pyrolytic graphite (HOPG) surface. Shape variations of the NTlike structures, caused by the modulation of scanning parameters, were observed and analyzed.
Keywords: STM; highly oriented pyrolytic graphite; nanotubelike structures
Scanning tunneling microscopy (STM) and spectroscopy (STS) are widely employed to obtain information on local electronic properties, surface geometry, and nanostructuresurface interactions.1 We report on STM studies of the NTL structures formed on HOPG surface. STM (OMICRON ultrahigh vacuum (UHV) system) measurements have been carried out in UHV at room temperature, in constantcurrent mode using tungsten tips. HOPG specimen has been prepared by cleaving (adhesive tape) in air, and annealed in UHV chamber at 900 °C. Naturallyformed NTL structures located on steps were formed by the spontaneous folding of the HOPG graphene sheets, or by STM tip manipulation.2 The NTL, presented in Fig. 1, is located at a HOPG step edge, with an apparent width of 1.77 nm. The height difference of the graphene layers between the right and left side of the NTL structure is 0.86 nm, quite close to 51 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 51–52. © 2006 Springer. Printed in the Netherlands.
52 the height of two graphene sheets. The image has been acquired with two different tunneling currents, corresponding to region A (0.1 nA) and B (0.2 nA), respectively. Region A corresponds to the typical appearance of the NTL structure with continuity of the distorted graphite lattice on the curved structure.
Figure 1. STM image (8u8 nm2) of NTL structure on the step edge of the HOPG graphene sheet. Scanning parameters: gap voltage +0.253 V and current 0.1 nA (region A) and 0.2 nA (region B).
Variation of tunneling parameters leads to dramatic changes of the NTL apparent shape. In region B, the NTL presents a large modification of the NTL curvature that becomes a large depression in the graphite structure. The maximum depth of the depression, with respect to the unmodified structure, is 0.4 nm. Such variation is not observed when the tip is moving from right side to left side of Fig.1, i.e., the backward direction. Hence the shape change behavior is asymmetric. The NTL recovers its original shape by simply restoring a lower current, without presenting any hysteresis in the apparent height variation. These observations suggest that the deformations may actually be induced by the modification of the tipsubstrate interaction strength, and that the NTL structure reacts in an elastic way to the STM tip. Although these features may support deformationbased mechanism of the shapechanges, further detailed analysis is required to elucidate the obtained experimental results. The results suggest a possible development of STM as a tool to investigate the mechanical properties of nanosystems. References 1. T. W. Odom, J. H. Hafner, Ch. M. Lieber, in: Carbon Nanotube: Synthesis, Structures, Properties and Applications, edited by M. S. Dresselhaus, G. Dresselhaus and Ph. Avouris (SpringerVerlag, Berlin, 2001) pp. 177215. 2. H.V. Roy, C. Kallinger, B. Marsen, and K. Sattler, J. Appl. Phys. 83, 46954699 (1998).
INFLUENCE OF CATALYST AND CARBON SOURCE ON THE SYNTHESIS OF CARBON NANOTUBES IN A SEMICONTINUOUS INJECTION CHEMICAL VAPOR DEPOSITION METHOD Z. E. HORVÁTH, A. A. KOÓS,* Z. VÉRTESY, L. TAPASZTÓ, Z. OSVÁTH, P. NEMES INCZE, L. P. BIRÓ Research Institute for Technical Physics and Materials Science, P.O. Box 49, H1525, Budapest, Hungary K. KERTÉSZ, Z. SÁRKÖZI, AL. DARABONT BabesBolyai University, Faculty of Physics, Kogãlniceanu 1, RO3400, ClujNapoca, Romania
Abstract. The injection chemical vapor deposition (CVD) method allows the semicontinuous production of pure multiwall carbon nanotubes (MWCNTs). In order to find the most efficient catalyst material and carbon source, we investigated the quality and quantity of carbon nanotubes, when different metallocenes (ferrocene, cobaltocene and nickelocene) and hydrocarbons (benzene, toluene, xylene, cyclohexane, cyclohexanone, nhexane, nheptane, noctane and npentane) are injected into the reaction furnace. The obtained samples were analyzed by Transmission Electron Microscopy (TEM). The highest yield and the best quality were obtained when a mixture of ferrocenenickelocene was used as catalyst and xylene as carbon source. The weight of purified carbon nanotubes was higher than 50% of the weight of catalyst material for xylene and reached the 10% for all investigated carbon sources.
Keywords: carbon nanotubes; chemical vapour deposition; electron microscopy
The major advantage of the injection CVD method is the insitu and continuous generation of catalytic particles throughout the entire growth cycle. Therefore, the injection CVD method does not need a catalyst synthesis step and can be scaled up for continuous or semicontinuous production. We investigated the quality and quantity of carbon nanotubes when different metallocenes and hydrocarbons were injected into the reaction furnace. *To whom correspondence should be addressed. Antal Koós; email: [email protected]
53 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 53–54. © 2006 Springer. Printed in the Netherlands.
54 We used the following reaction parameters: 6 g metallocene in 100 ml hydrocarbon catalyst concentration (except for the cases when the solubility of the metallocene in the hydrocarbon was too low), 1 ml/min solution flowrate, 875 °C furnace temperature and 500 l/h Ar (carrier gas) flowrate. The samples were purified using a 45% HNO3 aqueous solution.
Figure 1. Left: comparison of the weight of purified carbon nanotubes produced with different catalysts and carbon sources (black) and the weight of purified carbon nanotubes normalized to the weight of the catalyst material (grey). Right: TEM image of MWCNTs prepared using xylene as a carbon source.
Pure MWCNTs were grown. The MWCNTs were produced with maximum yield when ferrocene – nickelocene catalyst mixture was used and xylene was found to be the most efficient carbon source. For xylene, the weight of purified carbon nanotubes normalized to the weight of catalyst material was higher than 50%. Work supported by the Sapientia Research Programs Institute in the framework of Fellowships 822/2001 K/513/2003.03.25, and by the MTAMEH Grant "Nanogas" and by OTKA through Grant nr. T043685 and M 041689.
PECVD GROWTH OF CARBON NANOTUBES ALEXANDER MALESEVIC,* A. VANHULSEL Flemish Institute for Technological Research, Boeretang 200, 2400 Mol, Belgium C. VAN HAESENDONCK Solid State Physics, KULeuven, Celestijnenlaan 200 D, 3001 Heverlee, Belgium
Abstract. We report on the growth of carbon nanotubes by plasmaenhanced chemical vapor deposition (PECVD). We used acetylene and ammonia gas mixtures with a process pressure of 1 Torr for the growth of carbon nanotubes on nickel layers. The metal catalyst, just a few nanometers thick, was patterned on oxidized silicon substrates by electron beam lithography. Our aim was to develop a cheap mass production method to grow high quality, aligned carbon nanotubes. We used different plasma setups and scanned individual growth parameters in order to achieve that goal. In this work, we compare the different results obtained by direct current (DC) plasma deposition. Keywords: carbon nanotubes; plasmaenhanced chemical vapor deposition; DC plasma deposition; mass production
Whereas most groups concentrate on only one particular type of plasma for the growth of carbon nanotubes, we have built a unique setup which combines DC, radiofrequency (RF) and inductivelycoupled (ICP) plasma sources. A scheme of the setup is drawn in Fig. 1, left. In the case of a DC or RF bias, we apply a negative voltage to a resistively heated electrode and ignite plasma between the electrode and the grounded reactor walls. A dense RFICP source can be used in addition to or in combination with the DC or RF substrate biasing. In order to compare the results of different types of plasma, we started to grow carbon nanotubes with the DC plasma enhanced chemical vapor deposition (PECVD) technique. *To whom correspondence should be addressed. Alexander Malesevic, VITO Materialen, Boeretang 200, 2400 Mol; email: [email protected]
55 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 55–56. © 2006 Springer. Printed in the Netherlands.
56
Figure 1. Left: Schematic view of a PECVD setup. Right: SEM image of carbon nanotubes obtained with an acetylene to ammonia ratio of 25/100 sccm.
The samples consist of a 5 nm thick nickel layer that is deposited by a conventional molecular beam epitaxy (MBE) technique on a silicon (100) substrate with a silicon oxide buffer layer of 200 nm. The samples were placed on the electrode and the reactor is pumped down to a base pressure of 106 mBarr. The electrode was resistively heated up to 700°C in a continuous flow of 100 sccm ammonia at 1 mBarr. The samples were annealed at 700°C during 10 minutes and afterwards pretreated in ammonia plasma of 600 V during 10 minutes. The growth was initiated by allowing the carbon containing gas, diluted with ammonia, in the reactor for 20 minutes. As carbonaceous gas, we used acetylene, methane, propane and butadiene in combination with ammonia. We varied the ratio of the carbonaceous gas to ammonia from 15/100 to 33/100 sccm. The combination acetylene with ammonia was each time successful for the growth of carbon nanotubes, independent of the gas ratio. A typical result is presented in Fig. 1, right. Among the other gas mixtures, only the combination of methane with ammonia yielded carbon nanotubes.
CARBON NANOTUBES GROWTH AND ANCHORAGE TO CARBON FIBRES TH. DIKONIMOS MAKRIS,* R. GIORGI, N. LISI, E. SALERNITANO ENEA Casaccia Research Centre, 00060 Rome, Italy M. F. DE RICCARDIS, D. CARBONE ENEA Brindisi Research Centre, 72100 Brindisi, Italy
Abstract. A novel material made of carbon fabrics with a uniform threedimensional (3D) distribution of carbon nanotubes (CNTs) on the surface was synthesized by an electrochemical deposition (ELD) process. Keywords: carbon nanotubes; electrochemical deposition
The CNT exceptional tensile strength and modulus are deemed to improve the tensile properties of a polymeric matrix composite. In order to fully utilise the mechanical properties and the lightness of CNT, a sufficient volume fraction should be dispersed into the matrix.1 The problems are connected with the CNT aggregation and floating. The 3D distribution achieved by coating fabrics of carbon fibres with CNT could provide a way for dispersing the CNT across the composite. Moreover, CNT could further enhance the mechanical properties of composite materials traditionally reinforced with carbon fibres by enhancing the interfacial bonding between polymer and carbon fibre acting as fillers of micropores.2 ELD process was used to deposit Ni catalyst on PAN and Pitch based carbon fibres. Different deposition conditions were tested in order to optimise the dimension, morphology, and density of the Ni clusters.3,4 A Hot Filament assisted CVD reactor was used for the CNT growth.5,6 On both PAN and Pitch based carbon fibres, independently of the cluster size, density and morphology, clean CNT with low impurity content and rather smooth walls were grown.6 A high density of CNT was obtained on carbon fibres fabric (see Fig. 1, left), even at the critical cross over regions, as shown in Fig. 1, center. This result indicates *To whom correspondence [email protected]
should
be
addressed.
57 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 57–58. © 2006 Springer. Printed in the Netherlands.
Th.
Dikonimos
Makris;
email:
58 that all fabric fibres were coated by the catalyst giving rise to such a dense CNT network. Mechanical test usually performed to measure the adhesion of thin films on a substrate are not reliable in this case due to the peculiar nature of the substrate and the coating itself. Pseudomechanical tests, based on the idea of exposing the threecomponent material (carbon fibre plus metallic cluster plus carbon nanotubes) to stressing environments were performed.7 Samples were treated by immersion, magnetic stirring, centrifugation and ultrasonic bath in four different liquids (deionised water, ethanol, nbutanol, and acetone). After the treatment, the Ni clusters remained anchored to the fibres. As illustrated in Fig. 1, right, for the ultrasonic bath test in acetone, CNT were still present with high density on the fibre surface.
Figure 1. Left and center: high density CNT grown on carbon fibres fabrics. Right: carbon fibres after ultrasonic bath in acetone for 5 minutes.
References 1. 2. 3. 4. 5. 6. 7.
C. Bower, R. Rosen, L. Jin, J. Han, and Q. Zhou, Deformation of carbon nanotubes in nanotubepolymer composites, Appl.Phys.Lett.74(22), 33173319 (1999). E. T. Thostenson, W. Z. Li, D. Z. Wang, Z. F. Ren, and T. W. Chou, Carbon nanotube/carbon fibre hybrid multiscale composites, Journal of Applied Physics 91, 60346037 (2002). E. Gómez, R. Pollina, and E. Vallés, Morphology and structure of nickel nuclei as function of the conditions of electrodeposition, J. of Electroanalytical Chemistry 397, 111118 (1995). E. Gómez, R. Pollina, and E. Vallés, Nickel electrodeposition on different metallic substrates, J. of Electroanalytical Chemistry 386, 4546 (1995). Th. Dikonimos Makris, R. Giorgi, N. Lisi, L. Pilloni, E. Salernitano, F. Sarto, and M. Alvisi, Carbon nanotubes growth by HFCVD: effect of the process parameters and catalyst preparation, Diamond and Related Materials 13, 305310 (2004). Th. Dikonimos Makris, R. Giorgi, N. Lisi, L. Pilloni, E. Salernitano, and F. De Ricardis, Carbon nanotubes growth on PAN and Pitch based carbon fibres by HFCVD, Fullerenes, Nanotubes and Carbon Nanostructures 13(1), 383392 (2005). M. F. De Riccardis, D. Carbone, Th. Dikonimos Makris, R. Giorgi, N. Lisi, and E. Salernitano, Anchorage of carbon nanotubes grown on carbon fibres, Carbon (2005) (submitted).
CVD SYNTHESIS OF CARBON NANOTUBES ON DIFFERENT SUBSTRATES TH. DIKONIMOS MAKRIS, L. GIORGI, R. GIORGI, N. LISI, E. SALERNITANO ENEA Casaccia Research Centre, 00060 Rome, Italy M. ALVISI, A. RIZZO ENEA Brindisi Research Centre, 72100 Brindisi, Italy
Abstract. Carbon nanotubes (CNTs) were grown using three different chemical vapor deposition (CVD) processes. Optimized conditions were studied. Keywords: carbon nanotubes; chemical vapor deposition process
CNTs were grown on differently supported Ni catalytic nanoparticles on flat and bulk substrates using H2 and CH4 as precursors. The different behavior of the same metal catalyst in the presence of the same precursor varying the gas activation by different energy sources, using Hot Filament, Plasma Enhanced (PE), and pure Thermal CVD processes, was studied. By properly choosing the process parameters, dense CNTs were grown by HFCVD on 3 nm Ni thin film deposited by Evaporation and Radiofrequency (RF) Sputtering onto flat Si substrates coated with an intermediate SiO2 layer3 (Fig. 1a). Prior to the growth, the samples were heated in H2 atmosphere and the Ni clusters distribution shown in Fig. 2a was obtained. No CNT growth was obtained on the same sample in the thermal CVD reactor, but only cluster coalescence was observed (Fig. 2b). CNT grew sparsely in the PE CVD reactor. The different results were ascribed to: i) the catalytic decomposition of the precursors was more efficient where an additional activation source was present; ii) weak interaction between the Ni cluster and the SiO2 substrate could favour cluster coalescence, which was the dominant effect in the thermal CVD process. Flat Al2O3 substrate coated by a 3nm Ni film, deposited by RF Sputtering, were subjected to the same clustering and CNT growth process in Thermal and PE CVD. The thermal process failed as in the previous case giving rise to *To whom correspondence [email protected]
should
be
addressed:
59 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 59–60. © 2006 Springer. Printed in the Netherlands.
Th.
Dikonimos
Makris;
email:
60 aggregation of Ni clusters distributed on alumina grain surfaces. On the opposite, in the PE reactor, a dense distribution of Ni nanoclusters was obtained (Fig. 2c), followed by a dense CNT growth (Fig. 1b). The commercial catalyst, consisting of Ni nanoparticles supported on alumina microparticles, was tested in the thermal CVD process for CNT growth (Fig. 1c). This catalyst exhibited a high productivity and a yield of 15mgC/mgNi was obtained.4 The reason for this could be attributed to the higher metal dispersion on highly porous material. Moreover, the stronger interaction existing between the catalyst and the support prevented metal particles from aggregating and forming unwanted large particles. It could not be excluded that even an electronic interaction existed and it was responsible for the lowering of the catalyst activation barrier.
Figure 1. CNT growth on Ni/SiO2/Si flat sample in HFCVD (a), on Ni/Al2O3 flat sample in PECVD (b) and on Ni/Al2O3 pellet in Thermal CVD (c).
Figure 2. Cluster distribution on Ni/SiO2/Si flat sample in HFCVD (a) and Thermal CVD (b) and on Ni/Al2O3 flat sample in PECVD.
References 1.
2.
3.
4.
S. Hoffmann, B. Kleinsorge, C. Ducati, A. C. Ferrari, and J. Robertson, Lowtemperature plasma enhanced chemical vapour deposition of carbon nanotubes, Diamond and Related Materials 13, 11711176 (2004). K. Hernadi, Z. Konya, A. Siske, J. Kiss, A. Oszko, J. B. Nagy, and I. Kiricsi, On the role of the catalyst support and their interaction in synthesis of carbon nanotubes by CCVD, Mater. Chem.Phys. 77, 536541 (2002). Th. Dikonimos, R. Giorgi, N. Lisi, E. Salernitano, L. Pilloni, M. Alvisi, and F. Sarto, Carbon nanotubes growth by HFCVD: effect of the process parameters and catalyst preparation, Diamond and Related Materials 13, 305310 (2004). Th. Dikonimos, L. Giorgi, R. Giorgi, N. Lisi, and E. Salernitano, CNT growth on alumina supported nickel catalyst by thermal CVD, Diamond and Related Materials 14, 8159 (2005).
INFLUENCE OF THE SUBSTRATE TYPES AND TREATMENTS ON CARBON NANOTUBE GROWTH BY CHEMICAL VAPOR DEPOSITION WITH NICKEL CATALYST R. RIZZOLI, R. ANGELUCCI, S. GUERRI, F. CORTICELLI CNR  IMM Bologna, via P. Gobetti 101, 40129 Bologna, Italy M. CUFFIANI, G. VERONESE Department of Physics, University of Bologna, viale Berti Pichat 6/2, 40127 Bologna, Italy
Abstract. Our investigation aimed at the development of a process capable of producing well ordered, vertically aligned carbon nanotubes (CNTs) arrays, with reproducible properties, for applications such as a position particle detector and a cold cathode emitter for storage devices. The results on catalyst nanoparticles formation from a thin Ni film evaporated on SiO2 and Si3N4 substrates are presented. The substratecatalyst layers have been processed in several gaseous atmospheres and in the temperature range 700900°C in order to obtain the most appropriate morphology, size and density of the nanoparticles for the subsequent CNTs growth. The smallest nanoparticles have been obtained on the SiO2 substrate in H2 atmosphere and at the lowest temperature 700°C. However, the best vertically aligned and well graphitized CNTs resulted from the NH3 annealing process followed by the deposition of CNTs at 900°C in C2H2 and NH3.
Keywords: Carbon nanotubes; Ni catalyst; chemical vapor deposition synthesis
Ni thin films, 2 nm thick, were deposited by egun evaporation on SiO2 / Si and Si3N4 /Si substrates. The Ni nanoparticles (NPs) formation was studied varying the gas atmosphere (Ar, NH3 and H2) and the temperature (in the range 700°C900°C) applied during the warmup and annealing steps prior to the CNTs deposition step. Vertically aligned CNTs were grown by atmospheric pressure chemical vapor deposition in C2H2 + NH3 gas mixtures at temperatures in the range 750900°C.
61 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 61–62. © 2006 Springer. Printed in the Netherlands.
62 The investigation on the formation of Ni NPs, starting from a 2 nm thick Ni film, showed that the NP dimensions are mainly affected by the nature of the substrate layer. The NPs average diameter is smaller on SiO2 than on Si3N4 substrates. The use of a reducing (H2 or NH3) or inert (Ar) gas atmosphere during the NPs thermal treatment had a strong impact on their properties: size, morphology and chemical composition of subsurface layers both of the NPs and of the underlying substrate. For each gas atmosphere, with increasing the annealing temperature in the range 700°  900°C on Si3N4 substrates, the NPs size decreases, while on SiO2, it remains almost unchanged. The smallest NPs, shown in the SEM image of Fig. 1, were obtained on SiO2 after annealing in H2 at 700°C (scale bar on the bottom left side corresponds to 100 nm).
Figure 1. SEM images of Ni NPs on a SiO2 substrate after annealing at 700°C in H2 .
However, the best multiwalled CNTs, vertically aligned and well graphitized, resulted from the NH3 annealing process followed by the deposition of CNTs in 5% C2H2 in NH3 at 900°C, on a SiO2 substrate. SEM and HRTEM images of these CNTs are displayed in Fig. 2.
Figure 2. Left: crosssection SEM image of CNTs grown in C2H2 + NH3 at 900°C on Ni NPs formed in NH3; Right: HRTEM micrograph of one of these MWNTs.
NON CATALYTIC CVD GROWTH OF 2DALIGNED CARBON NANOTUBES NADEZDA I. MAKSIMOVA, JORG ENGSTLER, JORG J. SCHNEIDER* Dept of Chemistry, EduardZintl Inst. of Inorganic and Physical Chemistry, Tech. University, 64287 Darmstadt, Germany
Abstract. Aligned carbon nanotubes (CNTs) were grown in the channels of porous alumina membranes via a noncatalytic chemical vapor deposition (CVD) method. The morphology of the samples was studied.
Keywords: carbon nanotubes; noncatalytic chemical vapor deposition method
Carbon nanotubes (CNTs) produced by catalytic methods contain metal nanoparticles that cannot be removed completely and may have negative effect on the performance of the CNTs for microelectronic applications.1 Herein, we present an optimized synthesis procedure for wellaligned CNT arrays. It includes a new postprocess purification approach to remove carbonaceous byproducts. Aligned CNT arrays were grown in the tubular channels of porous anodic alumina oxide (PAOX) membranes (Anodisc 25, Whatman) with different pore diameters at 900qC by a CVD method based on propylene decomposition.2 The morphology of the samples was examined using scanning electron microscopy (Philips XL30 FEG). In the noncatalytic CVD process, the carbon atoms or reactive Cfragments within a PAOX template are directly deposited on the surface of PAOX membrane (Fig. 1, left). In this case, the growth of CNTs is mainly controlled by the shapes of the membrane pores and the catalytic action of the PAOX template.3 Besides the CNT growth inside the pores (Fig. 1, center), there is deposition of carbon byproducts on the PAOX membrane surface that cannot be prevented. This carbon deposit has to be removed before any application of CNTs could be considered. The presence of at least three carbon layers on the PAOX surface, i.e., a thick amorphous carbon overgrowth layer, a *To whom correspondence should be addressed: Jorg J. Schneider
63 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 63–64. © 2006 Springer. Printed in the Netherlands.
64 layer of high molecular hydrocarbons (tar, pitch, resin, etc.) penetrated on the PAOX surface, and a surface graphitic overlayer, was detected (Fig. 1, right). This allowed us to develop a controllable removal of PAOX template: (i) the surface carbon layer was oxidized in a controlled fashion at 550°C in air; (ii) high molecular hydrocarbon byproducts were removed by boiling in benzene for 3 hours; (iii) the PAOX template was removed via dissolution in concentrated HF or H3PO4 during 4 hours. After these procedures, freestanding CNTs were obtained, however, still stuck to each other and connected via a surface graphitic overlayer.2 The complex structural composition of the deposited carbon products is an evidence for the complicated mechanism of the noncatalytic metalfree CVD process containing only one precursor. (b)
precursor gas flow
Membrane
Figure 1. Scheme of the mechanism of CNT growth within the pores of a PAOX membrane by CVD with (left) and without catalyst (center). Right: crosssection of the PAOX membrane with CNTs inside the pores (a); scheme of carbon layers on the PAOX membrane surface (b).
In the present work, CNTs were produced within the pores of PAOX membranes by a non catalytic CVD method. The presence of three different carbon layers on the PAOX membrane surface after CVD process was proven and the threestep purification procedure was developed to receive the freestanding aligned CNT arrays. We acknowledge support of the Volkswagen Foundation (grant no. 53200034).
References 1. 2. 3.
W. A. de Heer, A. Chatelain and D. Ugarte, A Carbon Nanotube FieldEmission Electron Source, Science 270, 11791180 (1995). J. J. Schneider, N. I. Maksimova and J. Engstler (unpublished). L. Zhi, J. Wu, J. Li, U. Kolb, and K. Müllen, Carbonization of disclike molecules in porous alumina membranes: Toward carbon nanotubes with controlled graphenelayer orientation, Angew. Chem. Int. Ed. 44, 21202123 (2005).
PYROLYTIC SYNTHESIS OF CARBON NANOTUBES ON Ni, Co, Fe/ɆɋɆ41 CATALYSTS KSENIIA KATOK,* SERGIY BRICHKA, VALENTYN TERTYKH, GENNADIY PRIKHOD’KO Institute of Surface Chemistry of National Academy of Sciences of Ukraine, 17 Naumov Str., 03680 Kyiv164, Ukraine
Abstract. The purpose of the accomplished work was to use the advantages (ordered porous structure and high surface area) of mesoporous ɆɋɆ41 silicas in the pyrolytic synthesis of carbon nanostructures. MCM41 matrices were used for chemisorption of volatile nickel, cobalt, and iron acetylacetonates with following metal reduction. The synthesis of carbon nanotubes was carried out by two routes: (a) ɆɋɆ41 treatment with Ni(acac)2 or Co(acac)2 vapor at 150qɋ, metal reduction with H2 at 450qɋ, C2H2 pyrolysis over the catalyst at 700qɋ; (b) ɆɋɆ41 modification with Co(acac)2 or Fe(acac)2, reduction of supported metal complexes with C2H2, and pyrolysis at 700qɋ (in situ reduction). The TEM data showed that carbon nanotubes were formed with external diameter 1035 nm for Ni, 4284 nm for Co, and 1424 nm for Fecontaining matrices. In the absence of metals, low yield of nanotubes (up to 2%) was detected. Keywords: mesoporous MCM41 silica; metal acetylacetonates; acetylene; pyrolysis; carbon nanotubes; in situ reduction
The unique properties of the carbon nanotubes and the prospects of their possible applications continue to attract the attention of the researchers. The pyrolysis of hydrocarbons on the surface of inorganic matrices is one of the possible routes of carbon nanotubes synthesis. In this regard, the ordered mesoporous matrices, such as ɆɋɆ41, are of great importance. The synthesis of carbon nanotubes was carried out by two routes: (a) ɆɋɆ41 treatment with Ni(acac)2 or Co(acac)2 vapour at 150qɋ, metal reduction with H2 at 450qɋ, C2H2 pyrolysis over the catalyst at 700qɋ; (b) ɆɋɆ41 *To whom correspondence should be addressed. Kseniia Katok, Institute of Surface Chemistry of National Academy of Sciences of Ukraine 17 Naumov Str., 03164 Kyiv, Ukraine; email: [email protected]
65 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 65–66. © 2006 Springer. Printed in the Netherlands.
66 modification with Co(acac)2 or Fe(acac)2, reduction of supported metal complexes with C2H2, and pyrolysis at 700qɋ (in situ reduction). During the pyrolysis of acetylene, a deposition of decomposition products of black color on silica matrices was observed. The synthesized composite was treated with a solution of hydrofluoric acid at ambient temperature for removal of the silica phases. As a result, insoluble carbon fraction was obtained. The carbon nanotubes were identified using a transmission electron microscope. Cetyltrimethylammonium bromide was used as a template for the synthesis of MCM41 silica with specific surface area905 m2g1 and total volume of pores 0.87 cm3g1. The average diameter of the pores, determined by the ȼJHmethod, was about 39 Å. The characterization of the mesoporous silica was performed by nitrogen adsorption measurements at 77 K. In order to provide the homogeneous catalyst deposition, the silicas were treated with volatile cobalt (II) nickel (II) and iron (II) acetylacetonates at moderate temperature. At these conditions, acetylacetonates of metals are chemisorbed on the silica surface due to reaction with the surface silanol groups with eliminating one of the ligands. The removal of the ligands from the metal complexes supported on the silica was studied in air by the thermogravimetric technique. The destruction of the residual ligands of acetylacetonates (after interaction with silanol groups) occurred at about 445qC. As a result, formation of complex oxides of cobalt and iron occurred on the silica surface. The pyrolysis of acetylene on the surface of ɆɋɆ41 silicas containing metallic catalysts was studied. The application of matrices with supported metals as catalysts enabled us to attain rather high yields of carbon nanostructures (2030% of the process product). The electron microscopy data provided evidence for the formation of nanotubes with an external diameter of 4284 nm in the case of cobalt catalyst and 1424 nm in the case of iron catalyst. The formation of carbon fibers with diameters in the interval 80110 nm for nickel catalyst was also observed. Here it is appropriate to mention that in the absence of a catalyst on such matrices, small yields (up to 2%) of carbon nanotubes were attained.
A GRAND CANONICAL MONTE CARLO SIMULATION STUDY OF CARBON STRUCTURAL AND ADSORPTION PROPERTIES OF INZEOLITE TEMPLATED CARBON NANOSTRUCTURES TH. J. ROUSSEL, C. BICHARA, R. J. M. PELLENQ Centre de Recherche sur la Matière Condensée et Nanosciences – CNRS, Campus de Luminy, 13288 Marseille CEDEX 9, France
Abstract. We used the GrandCanonical MonteCarlo technique (GCMC) to simulate the vapor deposition of carbon in the porosity of various zeolitic nanopores (Silicalite, AlPO45, faujasite). The carboncarbon interactions were described within a TightBinding formalism (TB) and the carbonmatrix interactions were assumed to be physisorption. Depending on the pore size and topology, various carbon structures can be obtained. Keywords: adsorption; Grand Canonical MonteCarlo simulation; carbon nanotubes; zeolite; FAU; AlPO45; TightBinding
Zeolites are now involved in the process of manufacturing metallic or semiconductor nanostructures (clusters, nanowires and nanotubes). The basic idea is the deposition of an element from vapour or liquid phase inside the porosity of such crystalline materials that have a narrow pore size distribution. Then, nanostructures of the deposited element can be obtained after matrix removal. In our simulations, the carboncarbon interactions are described in the tight binding approximation (TB) that is a parameterized version of the Hückel theory. The interaction of carbon with Si, Al, P and O atoms comprising the different zeolite frameworks considered in this work, is assumed to remain weak in the physisorption energy range using a PNTrAZ potential.1 Experimental results for carbon adsorption in zeolites can be found in a series of papers2 where the synthesis of ultrasmall Single Wall Carbon Nanotubes (SWNTs) in the 0.73 nm channels of zeolite AlPO45 was described. We show (Fig.1, left) that the narrowest single wall nanotube can be synthesized in the cylindrical channel of AlPO45 zeolite (pore diameter 0.7 nm) in agreement with the experimental work of Tang et al.2 We further show 67 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 67–68. © 2006 Springer. Printed in the Netherlands.
68 that this ultra small nanotube (0.4 nm diameter) is defective but stable after template removal. Another example is the case of adsorption in zeolite silicalite.3 This purely siliceous zeolite possesses two types of intercrossing channels of 0.55 nm in diameter. We also demonstrate that such zeolite with smaller diameter of the channels does not allow obtaining nanotubes but a mesh of intercrossing carbon chains (Fig. 1, middle). Carbon adsorption was also attempted in the porosity of faujasite:4 a zeolite that has a porous network made of cages of 1 nm in diameter, interconnected with 0.7 nm large windows. The pore topology of faujasite (made of tetrahedrally coordinated spherical cavities) allows making a highly porous ordered carbon material (CNaY) adopting a diamondlike structure that was subsequently tested for H2 storage (Fig. 1, right).
Figure 1. Left: SWNTs in AlPO45; Middle: carbon in silicalite; Right: NaY replica (inset: unit cell).
The GCMC results show that carbon adsorption in AlPO45 allows the formation of an ultra small defective nanotube of 0.4 nm diameter. Our simulations show that this tube is stable upon matrix removal. In the case of silicalite, we only obtain a mesh of intercrossing carbon chains but no aromatic carbon structures indicating that the smallest host cavity size for growing aromatic carbon nanostructures is around 0.7 nm. The CNaY material is a porous carbon structure, which is very stable upon matrix removal and keeps the topology of the template, and is a possible gas storage device for hydrogen.
References 1. 2. 3. 4.
R. J.M. Pellenq and D. Nicholson, Intermolecular potential function for the physical adsorption of rare gases in silicalite1, J. Phys. Chem. 98, 1333913349 (1994). N. Wang, Z. K. Tang, G. D. Li, and J. S. Li, Singlewalled 4 Angstrom Carbon Nanotube Array, Nature 408, 5051 (2000). C. Bichara, R. J.M. Pellenq, and J.Y. Raty J.Y., Adsorption of selenium wires in silicalite1 zeolite: a first order transition in a microporous system, Phys. Rev. Lett. 89, 1610 (2002). T. Kyotani, T. Nagai, S. Inoue, and A. Tomita, Formation of New Type of Porous Carbon by Carbonization in Zeolite Nanochannels, Chem. Mater. 9, 609615 (1997).
Part II. Vibrational properties and optical spectroscopies
VIBRATIONAL AND RELATED PROPERTIES OF CARBON NANOTUBES VALENTIN N. POPOV* Faculty of Physics, University of Sofia, 1164 Sofia, Bulgaria PHILIPPE LAMBIN Facultés Universitaires NotreDame de la Paix, 5000 Namur, Belgium
Abstract. The symmetryadapted approach to the study of the vibrational and related properties of carbon nanotubes is presented. The usually very large number of carbon pairs in the unit cell of the nanotubes, that hinders most of the microscopic studies, is conveniently handled in this approach by using the screw symmetry of the nanotubes and a twoatom unit cell. This allows the systematic simulation of various properties (vibrational, mechanical, thermal, electronic, optical, dielectric, etc.) of all nanotubes of practical interest: The application of symmetryadapted models to the study of some of these properties is illustrated in this review in two cases: a forceconstant approach and a tightbinding approach.
Keywords: phonons; phonon dispersion; forceconstant model; tightbinding model
1. Introduction The discovery of the carbon nanotubular structures in 1991 led to an avalanche of theoretical and experimental work motivated by their unusual properties and the possibility for their industrial application.1 Nowadays, carbon nanotubes are components of composite materials, various fieldeffect devices, fieldemitters in displays, hydrogen storage devices, gas sensors, etc. The theoretical study of the nanotubes is based on the use of an idealized structure: a uniform cylinder or an atomic structure with onedimensional periodicity. The practical ________ * To whom correspondence should be addressed. Valentin N. Popov, Faculty of Physics, University of Sofia, 5 J. Bourchier Blvd., 1164 Sofia, Bulgaria; email: [email protected]
69 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 69–88. © 2006 Springer. Printed in the Netherlands.
70 calculations in the latter case face an often insurmountable obstacle arising from the necessity to deal with a large number of carbon atoms. This hindrance can be overcome considering a usually neglected symmetry of the ideal nanotubes, the screw symmetry. In this review, we show how the explicit use of the screw symmetry allows the easy computational handling of practically all existing nanotube types. This is illustrated on the example of the lattice dynamics of nanotubes developed within a forceconstant approach (Sec. 3) and a tightbinding approach (Sec. 4). The theoretical part is appended with results of calculations, which are compared to available experimental data. 2. The nanotube structure and screw symmetry The ideal singlewalled carbon nanotube can be viewed as obtained by rolling up an infinite graphite sheet (graphene) into a seamless cylinder leading to coincidence of the lattice point O at the origin and another one A defined by the chiral vector Ch = (L1,L2) (see Fig. 1). Each tube is uniquely specified by the pair of integer numbers (L1,L2) or by its radius R and chiral angle ș. The latter is defined as the angle between the chiral vector Ch and the nearest zigzag of carbon–carbon bonds with values in the interval 0 ș ʌ/6. The tubes are called achiral for ș = 0 (“zigzag” type) and ș = ʌ/6 (“armchair” type), and chiral for ș 0, ʌ/6. Every nanotube has a translational symmetry with primitive translation vector T = (N1,N2) (see Fig. 1) where N1 L1 2 L2 / d and N 2 2 L1 L2 / d . Here d is the greatest common divisor of 2L1 + L2 and L1 + 2L2. The total number of atomic pairs in the unit cell is N N1 L2 N 2 L1. All the carbon atoms of a tube can be reproduced by use of two different screw operators2,3 (see Fig. 2). The screw operator {St} rotates the position vector of an atom at an angle ĳ about the tube axis and translates it at a vector t along the same axis. Thus, the equilibrium position vector R(lk) of the kth atom of the lth atomic pair of the tube is obtained from R(k) Ł R(0k) by using of two screw operators {S1t1} and {S2t2}
R (lk ) {S1 t1}l1 {S2 t 2 }l2 R (k ) S1l1 S2l2 R (k ) l1t1 l2 t 2 .
(1)
Here l = (l1,l2), l1 and l2 are integer numbers labeling the atomic pair, and k = 1, 2 labels the atoms in the pair. It is convenient to adopt the compact notation S (l ) S1l1 S 2l2 and t (l ) l1t1 l2 t 2 , and rewrite Eq. 1 as R (lk ) S (l )R ( k ) t (l ).
71
Figure 1. The honeycomb lattice of graphene. The lattice points O and A define the chiral vector Ch and the points O and B define the primitive translation vector T of the tube. The rectangle formed by the two vectors defines the unit cell of the tube. The figure corresponds to tube (4,2) with Ch = (4,2) and T = (4,–5).
Figure 2. A front view of a nanotube. The closed dashed line is the circumference and the straight dashed line is the axis of the tube. The nanotube can be constructed by mapping of the two atoms of the unit cell (depicted by empty symbols) unto the entire cylindrical surface using two different screw operators. 2
72 This "rolled" nanotube structure can be relaxed within tightbinding or abinitio models. In the relaxation procedure, one usually imposes the following constraints: preservation of the translational symmetry of the tubes, keeping all atoms on one and the same cylindrical surface. As parameters of the relaxation one can choose the tube radius, the primitive translation of the tube, and the angle of rotation and the translation of the screw operation for mapping the first atom of the twoatom unit cell unto the second one. The relaxed structure can be used to obtain refined values of various quantities. 3. Lattice dynamics of carbon nanotubes: forceconstant approach The symmetryadapted lattice dynamical model for nanotubes can be constructed as a Born model of the lattice dynamics based on a twoatom unit cell. In the adiabatic approximation, the atomic motion is conveniently decoupled from the electronic one. For small displacements u(lk) of the atoms from their equilibrium positions, it is customary to use the harmonic approximation and represent the atomic Hamiltonian as a quadratic form of u(lk). The equations of motion of the atoms are then readily derived in the form
mk uD (lk ) ¦ )DE (lk , l ' k ')uE (l ' k ').
(2)
l 'k 'E
Here mk are the atomic masses and ĭĮȕ(lk,l'k') are the force constants. The screw symmetry of the nanotube suggests searching for a solution of the type uD (lk )
1 mk
¦E SDE (l)eE (k q) exp i q l Zt ,
(3)
which represents a wave with wavevector q = (q1,q2) and angular frequency Ȧ(q). The atomic displacements remain unchanged under integer number tube translations at distance T and integer number rotations at an angle of 2ʌ. These two conditions lead to the following constraints on the wavevector components
q1 L1 q2 L2 q1 N1 q2 N 2
2S l ,
(4)
q.
(5)
Here l = 0, 1,..., N1 and S d q S . Using Eqs. 4 and 5, the atomic displacement Eq. 3 can be written as uD (lk )
1 mk
¦E SDE (l)eE (k ql ) exp i D (l)l z (l)q Zt ,
(6)
73 Here
D (l ) 2S (l1 N 2 l2 N1 ) / N c and z (l ) ( L1l2 L2l1 ) / N c
are dimensionless coordinates of the origin of the lth cell along the circumference and along the tube axis, respectively. Substituting Eq. 6 in Eq. 2, we get a system of six linear equations of the form
Z 2 (ql )eD (k ql )
¦E DDE (kk ' ql )eE (k ' ql ),
(7)
k'
where the dynamical matrix is defined as DDE ( kk '  ql )
1 mk mk '
¦J )DJ (0k , l ' k ') SJE (l ') exp i D (l ')l z(l ') q .
(8)
l'
The eigenfrequencies Ȧ(ql) are solutions of the characteristic equation DDE ( kk ' ql ) Z 2 ( ql )G DE G kk '
0.
(9)
Substituting the solutions Ȧ2(qlj) in Eq. 7, one can obtain the corresponding eigenvectors eĮ(kqlj) (j = 1, 2,…, 6). For each q there are 6N vibrational eigenmodes (phonons) but the number of the different Ȧ2(qlj) can be lesser due to degeneracy. Using Eq. 8 it can be proven that D is Hermitian and therefore Ȧ2(qlj) are real and eĮ(kqlj) can be chosen orthonormal. Due to the explicit accounting for the screw symmetry of the tubes in the presented symmetryadapted scheme, the computation time for each q scales as 63N. This ensures great advantage for phonon calculations of tubes with very large N in comparison to the approach that does not use the screw symmetry where the computation time scales as (6N)3. Practically, all observable nanotubes can be handled with the presented lattice dynamical model with respect to the numerical stability of the computations as well as to the computer memory and CPU time. 3.1. PHONON DISPERSION, ZONECENTER PHONONS, AND RAMAN INTENSITY OF THE RBM
The phonon dispersion of two nanotubes, calculated using force constants of valence forcefield type,2,3 is shown in Fig. 3. The small number of force fields used does not allow for the precise modeling of the highfrequency branches. In particular, the "overbending" of the inplane bondstretching vibrations in
74 graphene cannot be reproduced well, which results in an incorrect order of the zonecenter tangential phonons in nanotubes. However, this does not affect much the lowfrequency branches in nanotubes. The phonons with wavevector q inside the Brillouin zone are waves along the tube with wavelength Ȝ = 2ʌ/q. The zonecenter phonons are atomic motions repeating in all unit cells along the tube. The atomic displacements for chiral tubes are not generally restricted to definite directions in space. However, the displacements for achiral tubes can be classified as radial (R), circumferential (C), and axial (A). The zonecenter phonons of chiral tubes belong to the following symmetry species of point groups DN (Ref. 4)
* =3A1 +3A 2 +3B1 +3B2 +6E1 +6E 2 +6E3 +...6E N/21 ,
(10)
where modes A1,2 and B1,2 correspond to l = 0 and N/2, and modes Ei correspond to l = i. The Ei modes have 2l nodes around the tube circumference. The A1,2 modes are nodeless and the B1,2 ones have N nodes. Armchair and zigzag tubes have additional symmetry elements and the modes are classified by the irreducible representations of point group D2nh.
Figure 3. The phonon dispersion of nanotubes (9,9) and (15,0) obtained using force constants of the valence force field type. Notice the large number of phonon branches: 108 and 180.
75 Among the various zonecenter phonons, some are infraredactive (A2 + 5E1 in chiral tubes, 3E1u in armchair tubes, and A2u + 2E1u in zigzag tubes), others are Ramanactive (3A1 + 5E1 + 6E2 in chiral tubes, 2A1g + 2E1g + 4E2g in armchair tubes, and 2A1g + 3E1g + 3E2g in zigzag tubes), and the rest are silent. The A1(g), E1(g), and E2(g) phonons are observed in the scattering configurations (xx + yy, zz), (xz, yz), and (xx  yy, xy), respectively, for z axis along the tube. (The used here Porto notation (x1x2) means incident light polarized along x1 axis and scattered light polarized along x2 axis.) Largest Raman signal is observed for parallel scattering configuration and it originates from the following A1(g) modes: x one mode with a uniform radial atomic displacement (socalled radialbreathing mode or RBM) with frequency following roughly the power law 230/d (in cm1; d is the tube diameter in nm); x one highfrequency mode of about 1580 cm1, which is purely circumferential in armchair tubes and purely axial in zigzag tubes (or two neither purely circumferential, nor purely axial modes in chiral tubes). The Raman signal due to E1(g) and E2(g) phonons is usually very weak. Apart from lines due to singlephonon processes, bands arising from more complex processes are often observed in the Raman spectra. An example of such band is the D band (i.e., disorder band) due to the presence of impurities and defects in the nanotubes.
Figure 4. Nonresonant Raman intensity of an isolated nanotube, a bundle of two nanotubes, and an infinite bundles of nanotubes (10,10) calculated within a bondpolarizability model.5,6
76 The Raman lines of the highfrequency A1(g) modes of the different tubes are closely situated and modified by electronphonon and electronimpurity interactions. The modeling of the latter ones also faces the problem of considering far enough neighbors in order to reproduce correctly the overbending of the phonon branches of graphene, from which these modes originate. On the other hand, the lines of the RBMs in the measured spectra are often wellseparated and can be used for structural characterization of the nanotube samples. In bundles of nanotubes and multiwalled nanotubes, there are many breathinglike phonon modes having different Raman line intensities.57. This is illustrated in Fig. 4, where the Raman intensity of such modes of an isolated and bundled nanotubes, calculated within a bondpolarizability model, is displayed. It is clearly seen that in bundles there are more than one breathinglike mode and that the intertube interactions generally upshifts their frequencies. It has to be noted that while the bondpolarizability model describes well the nonresonant spectra of most semiconductors, the Raman spectra of nanotubes are usually measured under resonant conditions and nonresonant spectra are rarely observed because of the low signal. Any realistic calculations of the spectra should take into account the electronic band structure of the tubes (see Sec. 3). 3.2. MECHANICAL AND THERMAL PROPERTIES
The force constants are invariant under infinitesimal translations along and perpendicular to the tube axis that leads to the translational sum rules and to three zonecenter zerofrequency modes.3 The infinitesimal rotation invariance condition imposed on the force constants gives rise to a rotational sum rule and to an additional zerofrequency mode. The former three modes are translations along and perpendicular the tube, and the latter mode is a rotation of the tube about its axis. These four modes belong to the four acoustic modes of the tube. The longitudinal and rotational acoustic branches have linear dispersion at the zone center. The corresponding acoustic wave velocities vL and vR can be associated with Young's and shear moduli of the tubes, Y and G, as vL Y / U and vR G / U with ȡ being the mass density. The transverse waves have quadratic dispersion at the zone center and zero group velocity there. The estimations of the two moduli within forceconstants, tightbinding, and abinitio models yield approximately the same value for the moduli per unit tube circumference (inplane moduli) of about 350 J/m2 and 150 J/m2, close to the values for graphite. The Poisson ratio is given by v (Y / 2 G ) / G and, for the calculated moduli, it is about 0.17. The use of different definitions for the crosssectional area of the nanotubes can yield quite different values for the
77 moduli. Thus, assuming a crosssection as a ring of width 0.34 nm, one obtains 1030 GPa and 440 GPa slightly dependent on the tube type. For a crosssection as a full disk with radius R, the moduli with decrease with the increase of the radius, etc. In the case of bundles of many tubes, the crosssection should be 2 chosen as a hexagon with area 3 2 R d / 2 with d being the intertube separation resulting again in decreasing of Y with increasing tube radius. In practice, in order to make use of the large moduli of the nanotubes, e.g., in composite materials, it is necessary to bind covalently the nanotubes in ropes and/or to the other constituents of these materials. The bulk modulus K of infinite bundles of nanotubes can be derived in a similar way as the elastic moduli of isolated tubes.8 For smallradius tubes, K = KW is determined by the intertube forces and increases with the increase of the tube radius. For largeradius tubes, K = KT is predominantly due to the tube elasticity and decreases with the increase of the radius. For intermediate radii, K can be expressed approximately as KWKT /(KW + KT). The calculated K from the slopes of the acoustic branches of infinite bundles of tubes agrees well with the results of the moleculardynamics simulations.9 The disagreement for R > 15 Å is due to the assumption of circular crosssection of the tubes in the former model while the intertube interactions polygonize the crosssection of largeradius tubes.
Figure 5. Bulk modulus K of infinite bundles of achiral SWNTs, calculated within a forceconstant model,8 in comparison to the one, derived in molecular dynamics simulations.9
78 The specific heat and thermal conductivity of carbon nanotubes are determined mainly by the phonons while the electronic contribution is usually negligible. The phonon specific heat of carbon nanotubes Cv can be derived using the phonon frequencies calculated within the latticedynamical model by means of the expression10
=Z / kBT exp =Z / kBT D(Z )dZ , kB ³ 2 ª¬exp =Z / k BT 1º¼ 0 f
Cv (T )
2
(22)
where D(Ȧ) is phonon density of states. At high temperature, Cv tends to the classical limit 2kB/m § 2078 mJ/gK (m is the atomic mass of carbon). At low temperature, Cv is determined by the dispersion of the lowestenergy acoustic branches. For onedimensional systems, these are the transverse acoustic ones with quadratic dispersion and, therefore, D(Z ) v 1/ Z and Cv v T . As it can be seen from Fig. 6, this behavior can be observed below about 1 K. This behavior is essentially different from that of graphene ( Cv v T ) and graphite and infinite bundles ( Cv v T 3 ). The lowtemperature thermal conductivity has the same temperature behavior as the specific heat. At high temperature, scattering from the tube boundaries and Umklapp processes become dominant.
Figure 6. Specific heat of SWNTs and ropes of SWNTs, calculated within a forceconstant model,10, in comparison with that of graphene and graphite, and available experimental data.11,12,13
79 4. Lattice dynamics of carbon nanotubes: tightbinding approach 4.1. THE SYMMETRYADAPTED TIGHTBINDING MODEL
The electronic band structure of a periodic structure is usually obtained in the independentelectron approximation solving the oneelectron Schrödinger equation ª = 2 2 º V (r ) »\ k (r ) « ¬ 2m ¼
Ek\ k (r ) ,
(10)
where m is the electron mass, V(r) is the effective periodic potential, ȥk(r) and Ek are the oneelectron wavefunction and energy depending on the wavevector k. This equation can be solved by representing ȥk(r) as a linear combination of basis functions ĳkr(r)
\ k (r )
¦c
kr
Mkr (r ).
(11)
r
In the tightbinding approach, the ĳ's are constructed from atomic orbitals centered at the atoms. Let us denote by Ȥr(R(l) – r) the rth atomic orbital centered at an atom with position vector R(l) in the lth unit cell. Similarly to the vibrational problem considered above, the screw symmetry of the tube allows one to solve the electronic problem working with a twoatom unit cell.14 In this case, Bloch's condition for the basis functions ĳ is satisfied for the following linear combination of Ȥ's
Mkr (r )
1 Nc
¦e
ik l
Trr ' (l ) F r ' (R (l ) r ) .
(12)
lr '
where k is a twodimensional wavevector, k = (k1, k2), and Nc is the number of unit cells in the tube. Trr'(l) are appropriate rotation matrices rotating a given atomictype orbital to the same orientation with respect to the nanotube surface for all atoms. The wavevector components are subject to the following constraints arising from the rotational and translational symmetry of the tube k1 L1 k2 L2 k1 N1 k2 N 2
2S l ,
(13)
k.
(14)
Here k is the onedimensional wavevector of the tube (–ʌ k < ʌ) and the integer number l labels the electronic energy levels with a given k (l = 0,1, …, N–1). The other symbols are defined in Sec. 2. Using Eqs. 13 and 14, the expansion functions can be rewritten as
80 1 Nc
M klr (r )
¦ e D i
( l )l z ( l ) k
Trr ' F r ' (R (l ) r ).
(15)
l
After substitution of Eqs. 11 and 15 in Eq. 10, the electronic problem is transformed into the matrix eigenvalue problem
¦H
c
klrr ' klr '
r'
Ekl ¦ S klrr ' cklr ' ,
(16)
r'
where the matrix elements of the Hamiltonian Hklrr' and the overlap matrix elements Sklrr' are given by H klrr '
¦ e D
(l )l z (l ) k
¦ e D
(l )l z (l ) k
i
H rr '' (l )Tr '' r ' (l ) ,
(17)
S rr '' (l )Tr '' r ' (l ) .
(18)
lr ''
S klrr '
i
lr ''
Here Hkl(l) and Skl(l) are matrix elements of the Hamiltonian and the overlap matrix elements between Ȥ's. The set of linear algebraic equations Eq. 16 has nontrivial solutions for the coefficients c only for energies E which satisfy the characteristic equation H klrr ' Ekl Sklrr '
0.
(19)
The solutions of Eq. 19, Eklm, are the electronic energy levels; the energy bands are labeled by the composite index lm (m = 1, 2,…, 8 for 4 electrons per carbon atom). The corresponding eigenvectors cklmr are determined from Eq. 16. Like with phonons, the symmetryadapted approach for calculation of the electronic band structure of nanotubes has the important advantage of large reduction of the computational time. Indeed, the calculation of the band energies for a given k needs time scaling as the cube of the size of the twoatom eigenvalue problem (83 for 4 electrons per carbon atom) times the number of the twoatom unit cells N. On the other hand, the straightforward calculation of the same band energies will require time scaling as the cube of the size of the 2Natom eigenvalue problem, i.e., (8N)3. For structural relaxation of a nanotube one needs the expressions for the total energy and the forces acting on the atoms. The total energy of a nanotube (per unit cell) is given by E
occ
¦E klm
klm
1 ¦¦I (rij ) , 2 i j
(20)
where the first term is the band energy (the summation is over all occupied states) and the second term is the repulsive energy, consisting of repulsive pair
81 potentials I (r ) between pairs of nearest neighbors. The band contribution to the force in Į direction on the atom with a position vector R(0) is given by the HellmannFeynman theorem FD
occ
wEklm D ( 0)
¦ wR klm
occ
¦¦ c klm rr '
* klmr
w H klrr ' Eklm Sklrr ' wRD (0)
cklmr ' .
(21)
The repulsive contribution is the first derivative of the total repulsive energy with respect to the position vector R(0). According to the simplest tightbinding model with one valence electron per carbon atom (the ʌband TB model), tubes with residual of the division of L1 L2 by 3 equal to zero are metallic (Mtubes), and the remaining tubes with residual 1 and 2, denoted further with Mod1 and Mod2, are semiconducting (Stubes). The bandstructure calculations within more sophisticated models show that only armchair tubes are strictly metallic, while the remaining Mtubes are tinygap semiconductors. The results of the calculations of the band structure and the electronic density of states (DOS) for two SWNTs within a nonorthogonal tightbinding model14,15 (NTB model) are shown in Fig. 7. Tube (10,10) is metallic with small finite density of states at the Fermi energy defined as zero. Tube (17,0) is semiconducting with a small energy gap of about 0.6 eV. The DOS spikes have characteristic 1/ E singularity at the optical transitions.
Figure 7. Band structure and electronic density of states of SWNTs (10,10) and (17,0). Notice the characteristic 1/ E spikes in the DOS.
82 4.2. DIELECTRIC PROPERTIES
The imaginary part of the dielectric function in the randomphase approximation is given by14 4S 2 e 2 m 2Z 2
H 2 (Z )
2
¦ 2S ³ dk
pkl ' cklv , P G Ekl ' c Eklv =Z , 2
(22)
l ' clv
where ƫȦ is the photon energy, e is the elementary charge, and m is the electron mass. The sum is over all occupied (v) and unoccupied (c) states. pkl'cklv,µ is the matrix element of the component of the momentum operator in the direction µ of the light polarization pkl ' cklv , P
³ dr\
* kl ' c
(r ) pˆ P\ klv (r ) .
(23)
Substituting Eqs. 11 and 15 in Eq. 23, one obtains the nonzero matrix elements pkl ' cklv , P
f ll ' P ¦ ckl* ' cr ' cklvr ¦ e rr '
i D ( l ) l z ( l ) k
pr ' r '', P (l )Trr '' (l ) ,
(24)
lr ''
where prr ', P (l )
³ drF (R(0) r) pˆ P F r
r'
(R '(l ) r ) .
(25)
For zaxis along the tube axis, the quantities f ll ', P are given by f ll ' x f ll ', y (G l ',l 1 G l ',l 1 ) / 2 and f ll ', z G ll ' . The latter two relations express the selection rules for allowed dipole optical transitions, namely, optical transitions are only allowed between states with the same l for parallel light polarization along the tube axis and between states with l and l' differing by 1 for parallel polarization, which is perpendicular to the tube axis. In the latter case, the electronic response is much weaker than in the former due to the depolarization effect. Therefore, much of the efforts have been directed to the calculation of the dielectric function for parallel polarization. The optical transition energies can be determined from the spikes of the imaginary part of the dielectric function calculated within a NTB model.14,15 The derived energies for parallel light polarization are shown in Fig. 8 for HiPco samples and for the most frequently used laser excitation energies. The points on the plot form strips belonging to either M or Stubes. With increasing the energy, one crosses strips of points for transitions 11 in Stubes (Mod1, Mod2), 22 in Stubes (Mod2 and Mod1), 11 and 22 in Mtubes (only transition 11 in armchair tubes), and so on. In each strip, the arrangement of the points show family patterns for L1 + 2L2 = integer number (connected with lines).
83
Figure 8. The resonance chart: transition energy vs tube radius. Only the part, containing the points that correspond to the HiPco samples and the most frequently used laser excitation energies, is shown.
4.3. THE DYNAMICAL MATRIX
The dynamical matrix of the nanotubes can be obtained from the expression of the total energy by considering the change of the energy arising from a given static lattice deformation16 (a phonon). The derivation is too involved to be placed here and the same holds true for the final expression of the dynamical matrix. Here, it suffies to provide results of the calculations for particular nanotubes. We note that this method is not applicable to armchair tubes because the dynamical matrix has a logarithmic singularity at the band crossing at the Fermi energy. The tinygap Mtubes require large number of kpoints to ensure convergence of the sum over the Brillouin zone in the expression of the
84 dynamical matrix. This problem does not normally appear for Stubes where a moderate number of kpoints is sufficient. The calculated phonon dispersion of graphene within the NTB model overestimates the measured one by about 10 %. The calculated frequencies, multiplied by 0.9, are shown in Fig. 9. It is clear that the overbending of the highestfrequency branches is well reproduced while the outofplane optical branch is now underestimated. Since we are rather interested in the former one, the underestimation for the latter one will not be important. As it was mentioned above, the RBM is of primary importance for the structural characterization of the samples. In Fig. 10, the calculated RBM frequency of all tubes in the radius range from 2 Å to 12 Å is shown in comparison with available abinitio results for strictly radial atomic displacements. The points follow a power law 1141/R (in cm1, R in Å). In the presence of impurities and structural disorder, there is an additional lightscattering mechanism, socalled double resonance mechanism. It is responsible for the specific Raman line broadening and the position shift with the laser excitation energy especially observed for the D (disorder), G, and G' bands. The review of this mechanism lies outside the scope of this paper.
Figure 9. Phonon dispersion of graphite calculated within a NTB model and scaled by a factor of 0.9 (Ref. 16) in comparison with available experimental data.17,18
85
Figure 10. Calculated RBM frequencies within a NTB model vs tube radius.16 The comparison with abinitio results19 reveals disagreement less than 5 cm1 (see inset).
4.4. THE RESONANT RAMAN INTENSITY
The Ramanscattering process can be described quantummechanically considering the system of the electrons and phonons of the tube and the photons of the electromagnetic radiation and their interactions.20 The most resonant Stokes process includes a) absorption of a photon (energy EL, polarization vector İL) with excitation of the electronic subsystem from the ground state with creation of an electronhole pair, b) scattering of the electron (hole) by a phonon (frequency Ȧo, polarization vector e), and c) annihilation of the electronhole pair with emission of a photon (energy ES = EL  ƫȦo, polarization vector İS) and return of the electronic subsystem to the ground state. The Raman intensity for the Stokes process per unit tube length (the socalled resonance Raman profile or RRP) is given by21 (up to a slightly tubedependent factor) 2
pcvS Dcv pcvL* 1 , I ( E L , Zo ) v ¦ L cv EL Ecv iJ cv ES Ecv iJ cv
(26)
where L is the length of the tube containing Nc unit cells, Ecv is the vertical separation between two states in the valence band v = klm and the conduction band c = klm', and Ȗcv is the excited state width. pL,Scv is the matrix element of the component of the momentum in the direction of the polarization vector İL,S. The electronphonon coupling matrix element Dcv is determined by the scalar
86 product of the derivative of Ecv with respect to the atomic displacement vector u and the phonon eigenvector e Dcv v
wE = ¦ eiJ wucv . 2mZo iJ iJ
(27)
Here m is the mass of the carbon atom and uiȖ is the Ȗcomponent of the displacement of the ith atom in the twoatom unit cell. The derivatives with respect to the atomic displacements for the considered phonon can be calculated using the HellmannFeynman theorem. The estimation of the Raman intensity can be simplified for wellseparated features of the RRPs within two approximation schemes. In the first scheme, one considers the matrix elements in Eq. 26 as kindependent and pulls them out of the summation. The remaining summation can be performed analytically. The obtained RRPs reproduce within 5% the results of the full calculations using Eq. 26. In the second scheme, one considers the matrix elements as kindependent and tubeindependent (i.e., constants) and evaluates the remaining sum analytically. Based on extensive calculations, it was shown that the tubeindependent scheme often leads to erroneous predictions of the Raman intensity and should therefore be avoided.21,22 This is illustrated in Fig. 11 in the case of three tubes with close radii.
Figure 11. RBM resonance Raman profiles of three different tubes with radius of about 6.7 Å: (17,0), (14,5), and (10,10). The results of the full calculations (Eq. 26) are compared to those of the frequently used tubeindependent numerator approximation.22
Once the RRPs are evaluated for all nanotube types present in the measured samples, the Raman spectra for different laser excitation energies can be simulated using the following expression21
87 I tot ( EL , Z )
¦ gI ( E
L
, Zo )
1
Z Zo
2
J o2
,
(28)
where g is the tube distribution function for the given sample, Ȗo is phonon HWHM, and the summation is carried out over all tube types. The usually adopted Gaussian distribution does not describe sufficiently well the tube types present in the samples. In this case, Eq. 28 can be fit to the measured spectra at several laser excitation energies to obtain the distribution function. The maximum Raman intensities can be plotted for better visualization as circles with size proportional to the logarithm of the intensity on the transition energy  tube radius plot. As seen in Fig. 8, the lower part of the points of each strip has larger intensity than the upper one. Thus, the points belonging to transitions 11 in Stubes (Mod1), 22 in Stubes (Mod2), 11 in Mtubes (Mod0) including transitions 11 in armchair tubes have larger intensity than their counterparts in the strips. This general behavior is corroborated by recent Raman data.23,24 5. Conclusions It is shown that the atomistic approach to the study of the vibrational, mechanical, thermal, electronic, optical, dielectric, etc., properties of ideal nanotubes can be successful when exploiting the screw symmetry of the nanotubes. The results of the calculations of some of these properties within the symmetryadapted approach are discussed in comparison with available data. ACKNOWLEDGEMENTS
V.N.P. was supported partly by the MarieCurie IntraEuropean Fellowship MEIFCT2003501080. Both authors were supported partly by the NATO CLG 980422. Work performed within the IUAP P5/01 Project "Quantumsize effects in nanostructured materials" funded by the Belgian Science Programming Services.
References 1. 2. 3.
Carbon nanotubes: Synthesis, Structure, Properties, and Applications, edited by M. S. Dresselhaus, G. Dresselhaus, and Ph. Avouris (SpringerVerlag, Berlin, 2001). V. N. Popov, V. E. Van Doren, and M. Balkanski, Lattice dynamics of singlewalled carbon nanotubes, Phys. Rev. B 59, 83558358 (1999). V. N. Popov, V. E. Van Doren, and M. Balkanski, Elastic properties of singlewalled carbon nanotubes, Phys. Rev. B 61, 30783084 (2000).
88 4. 5.
6. 7. 8. 9. 10. 11.
12. 13. 14.
15.
16. 17. 18. 19. 20. 21.
22.
23.
24.
O. E. Alon, Number of Raman and infraredactive vibrations in singlewalled carbon nanotubes, Phys. Rev. B 63, 201403/13 (2001). V. N. Popov and L. Henrard, Evidence for the existence of two breathinglike phonon modes in infinite bundles of singlewalled carbon nanotubes, Phys. Rev. B 63, 233407233410 (2001). L. Henrard, V. N. Popov, and A. Rubio, Influence of Packing on the Vibrational Properties of Infinite and Finite Bundles of Carbon Nanotubes, Phys. Rev. B 64, 205403/110 (2001). V. N. Popov and L. Henrard, Breathinglike phonon modes in multiwalled carbon nanotubes, Phys. Rev. B 65, 235415/16 (2002). V. N. Popov, V. E. Van Doren, and M. Balkanski, Elastic properties of crystals of singlewalled carbon nanotubes, Solid State Commun. 114, 395399 (2000). J. Tersoff and R. S. Ruoff, Structural Properties of a CarbonNanotube Crystal, Phys. Rev. Lett. 73, 676679 (1994). V. N. Popov, Lowtemperature specific heat of nanotube systems, Phys. Rev. B 66, 153408/14 (2002). A. Mizel, L. X. Benedict, M. L. Cohen, S. G. Louie, A. Zettl, N. K. Budraa, and W. P. Beyermann, Analysis of the lowtemperature specific heat of multiwalled carbon nanotubes and carbon nanotube ropes, Phys. Rev. B 60, 32643270 (1999) J. Hone, B. Batlogg, Z. Benes, A. T. Johnson, and J. E. Fischer, Quantized Phonon Spectrum of SingleWall Carbon Nanotubes, Science 289, 17301733 (2000). J. C. Lasjaunias, K. Biljakoviü, Z. Benes, J. E. Fischer, and P. Monceau, Lowtemperature specific heat of singlewall carbon nanotubes, Phys. Rev. B 65, 113409/14 (2002). V. N. Popov, Curvature effects on the structural, electronic and optical properties of isolated singlewalled carbon nanotubes within a symmetryadapted nonorthogonal tightbinding model, New J. Phys. 6, 17.117.17 (2004). V. N. Popov and L. Henrard, Comparative study of the optical properties of singlewalled carbon nanotubes within orthogonal and nonorthogonal tightbinding models, Phys. Rev. B 70, 115407/112 (2004). V. N. Popov (unpublished). T. Izawa, R. Souda, S. Otani, Y. Ishizawa, and C. Oshima, Bond softening in monolayer graphite formed on transitionmetal carbide surfaces, Phys. Rev. B 42, 1146911478 (1990). S. Siebentritt, R. Pues, K.H. Rieder, A. M. Shikin, Surface phonon dispersion in graphite and in a lanthanum graphite intercalation compound, Phys. Rev. B 55, 79277934 (1997). J. Kürti, V. Zólyomi, M. Kertesz, and G. Sun, The geometry and the radial breathing mode of carbon nanotubes: beyond the ideal behaviour, New J. Phys. 5, 125.1125.21 (2003). R. Loudon, Theory of the firstorder Raman effect in crystals, Proc. Roy. Soc. (London) 275, 218232 (1963). V. N. Popov, L. Henrard, and Ph. Lambin, Resonant Raman intensity of the radial breathing mode of singlewalled carbon nanotubes within a nonorthogonal tightbinding model, Nano Letters 4, 17951799 (2004). V. N. Popov, L. Henrard, and Ph. Lambin, Electronphonon and electronphoton interactions and the resonant Raman scattering from the radialbreathing mode of singlewalled carbon nanotubes, submitted to Phys. Rev. B (2005). C. Fantini, A. Jorio, M. Souza, M. S. Strano, M. S. Dresselhaus, and M. A. Pimenta, Optical Transition Energies for Carbon Nanotubes from Resonant Raman Spectroscopy: Environment and Temperature Effects, Phys. Rev. Lett. 93, 147406/14 (2004). H. Telg, J. Maultzsch, S. Reich, F. Hennrich, and C. Thomsen, Chirality Distribution and Transition Energies of Carbon Nanotubes, Phys. Rev. Lett. 93, 177401/14 (2004).
RAMAN SCATTERING OF CARBON NANOTUBES H. KUZMANY,* M. HULMAN, R. PFEIFFER, F. SIMON Fakultät für Physik der Universität Wien, Wien, Austria
Abstract. The present state of Raman scattering from carbon nanotubes is reviewed. In the first part of the presentation, the basic concepts of Raman scattering are elucidated with particular emphasis on the resonance scattering. The classical and the quantummechanical descriptions are presented and the basic experimental instrumentation and procedures are described. Special Raman techniques are discussed. Eventually, a short review on the electronic structure of singlewall carbon nanotubes (SWCNTs) is given. The second part of the presentation deals with Raman scattering from SWCNTs. The group theoretical analysis and the origin of the basic Raman lines are described. For the radial breathing mode, the observed quantum oscillations and the unusual strong Raman cross section are discussed. For the Gline, the resonance behavior and the response to doping are demonstrated and the calculated dependence of the line frequency on the tube diameter is summarized. For the Dline and for the G'line, the dispersion is demonstrated and its origin from a triple resonance mechanism is described. Finally, the response from pristine and doped peapods is elucidated. In the third part, most recent results are reported from Raman spectroscopy of doublewall carbon nanotubes (DWCNTs). The unusual narrow lines with widths down to 0.4 cm1 indicate clean room conditions for the growth process of the inner tubes. The (n,m) assignment of these lines and the high curvature effects are discussed. Results for DWCNTs, where the inner tubes are highly 13Csubstituted, are reported with respect to Raman and NMR spectroscopy. Eventually, it is demonstrated that the RBM Raman lines of the inner tubes cluster into groups of up to 14 lines where each member of the cluster represents a pair of innerouter tubes.
Keywords: Raman scattering; singlewall carbon nanotubes; peapods; doublewall carbon nanotubes; nano clean room; isotope effects *To whom correspondence should be addressed. Hans Kuzmany; email: [email protected]
89 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 89–120. © 2006 Springer. Printed in the Netherlands.
90 1. Preface Raman scattering has developed into one of the key analytical tools to investigate carbon nanotubes. This holds in particular as far as singlewall carbon nanotubes (SWCNTs) are concerned. There are several reasons for this. Firstly, the structure of the tubes is rather simple and many properties can be derived to a first approximation from the study of zonefolded planar graphene. Secondly, the Raman spectra of carbon phases are very well known and easy to record. Thirdly, for the SWCNTs, resonance conditions are easily met for lasers operating in the visible spectral range. As a matter of fact, under good resonance conditions, the Raman response from the tubes has the highest cross section of all materials known so far. Finally, many of the physical properties of the SWCNTs scale with the diameter and depend on the helicity of the tubes, like some of the Raman lines. Therefore the latter are appropriate for deriving information on such properties. In this series of lectures, three major subjects will be discussed: x
Fundamentals of Raman scattering
x
Raman scattering of singlewall carbon nanotubes
x
Raman scattering of doublewall carbon nanotubes
2. Fundamentals of Raman scattering Raman scattering in solids is the inelastic scattering of light from interaction with quasiparticles with an optical type of dispersion. This means that for these particles Z(q) 0 even for q = 0. This contrasts the Raman process to Brillouin scattering where interaction of light is with quasiparticles with acoustic dispersion such that Z(q) = 0 for q = 0. A general description of Raman scattering can be found in Refs. [15].
2.1. FORMAL AND MATHEMATICAL DESCRIPTION OF THE RAMAN PROCESS
The classical approach to the scattering from solids considers the modulation of the susceptibility FjlZ by phonons. It is expanded with respect to the normal coordinate Qk of phonon k. The first expansion coefficient is the Raman tensor. Thus we obtain for the polarizability (in As/m2) induced by the phonon k
PD j
F jl , k (Z , : k , t )H 0 E l (Z , t )
1 H 0 F jl , k (Z , : k )Qk E l 0 Cos (Z r : k )t 2
(1)
91 with
F jl ,k
wF jl wQk
.
Assuming that the induced polarizations act as the source of the evanescent radiation in the form of a Hertzian dipole, we obtain for the emitted Raman light and a 90o scattering geometry as depicted in Fig.1, after thermal averaging,
= (Z r : k ) 4 Vu F yx ,k (nk 1)V 2
) yx , :
32S 2 c0 : k 4
Ii
(2)
in [W/Ster], where Ii, Vu, V and nk = exp(ƫ:k/kBT) are the intensity of the incident light, the unit cell volume, the scattering volume, and the average Bose occupation number, respectively. From Eq. 2 we define the Raman cross section of the form
V
1 d) yx , VI i d :
(3)
where d : is a differential solid angle. In the quantum mechanical description, the incident photon excites an electronhole pair from a ground state 0² to an intermediate state i². This transition does not need energy conservation. In the intermediate state a phonon is generated (or absorbed), either from the excited electron or from the excited hole, by electronphonon interaction. This takes the system to another intermediate state f’², from where the electronhole pair can recombine to the final state f². This state is higher (lower) in energy as compared to the original ground state 0² by one phonon. Overall energy has to be conserved. In solids
Figure 1. Left: geometry for a 900 Raman scattering experiment. Incident and scattered light are either parallel or perpendicular polarized to the scattering plane; Center: quantummechanical states involved in the scattering process; Right: Feynman graphs for the HpA interaction (phonon scattering, top), the HAA interaction (electron scattering, center), and for a twophonon process (bottom).
92 wave vector conservation is also required. This means: k i k s r q where ki and ks are the wave vectors of the incident and scattered light, respectively, and q is the wave vector of the phonon. Accordingly, due to the small value of the wave vectors of visible light, in first order, only phonons with q ~ 0 (*point phonons) can contribute to the scattering process. For the quantitative description of the quantummechanical formulation, we can use the Feynman diagrams of Fig. 1 (right). For the onephonon Stokes process this yields K 2 f ,10
¦
Z 2 , f H er 0, f ²¢ 0, f ' H ep 0, i²¢ 0, i H er Z1 , 0 ( E1 Ei i=J )( E2 E f i=J )
i, f
(4)
Where Her and Hep are the electronradiation and the electronphonon interaction Hamiltonians, respectively, E1 and E2 are the energy of the incident and scattered light, Ei and Ef represent all transition energies from the ground state and to the final state involved, and J is an inverse life time constant. The absolute square of this expression is proportional to the Raman cross section. The four indices of the symbol on the left hand side of the equation represent the two states 0² and f² and the two phonons Zand Zinvolved If one or several transitions from Eq. 5 are close to a real eigenstate, the cross section will be strongly enhanced and is easier to calculate as it will be dominated by these transitions. This case is particularly important for solids where the transition energies often exhibit a continuum in the visible spectral range, known as the joint density of states gjds(H . In this case we can replace the sum in Eq. 4 by an integral and obtain for the Raman cross section
V (Z1 ) v
M f M ph M 0 g jds (H ) dH
³ (=Z H i=J )(=Z 1
M
2
2
H i=J )
2

(5)
g jds (H ) dH
³ (=Z H i=D )(=Z 1
2
H i=D )
where M stands for the product of the three matrix elements on the left side. 2.2. RAMAN SCATTERING, EXPERIMENTAL
On the experimental side, four basic components are needed for Raman scattering: a very narrow band light source, a light focusing and light collection optics, a monochromator system with very good stray light suppression and a very sensitive photon detector. A schematic of a Raman experiment is depicted in Fig. 2.
93
Figure 2. Experimental set up for a Raman experiment, consisting of several lasers for excitation, a light focusing and light collection optic, the sample, the monochromator and an appropriate detector.
As a light source almost exclusively lasers are used in the visible or near visible spectral range. A few mW of power are enough. Tunable lasers such as dyelasers or TiSapphire lasers have great advantage if resonance scattering is the goal. The focusing and light collection system is either a conventional bright optics or a microscope. In special cases, nearfield optics can be used. In this case, the spatial resolution goes down to about 40 nm. For the spectral analysis, usually triple monochromators are used which can be operated in an additive or in a subtractive mode. In the former, the spectral resolution is higher and allows for a spectral band pass down to 0.5 cm1 for red lasers. In the subtractive mode, the stray light suppression is much better which allows measuring Raman shifts down to the one wave number region. Recently, single monochromators have also come in use (confocal Raman systems) in combination with very sharp cutting notch filers. The advantage of such systems is their very high brightness but each laser needs its own notch filter. For the detection of the Raman light, almost exclusively nitrogen cooled, backthinned, and, if necessary, blue enhanced CCD detectors are used. Raman scattering has been widely applied to study crystalline and noncrystalline materials with respect to their structures, bonding, symmetries, morphologies, phase transitions, etc. From the many applications, two are selected here as they will be important for the following. Figure 3 depicts Raman spectra of various wellknown carbon phases. Diamond has cubic crystal symmetry and two atoms per unit cell. Therefore, it has only one, but threefold degenerated optical mode. Graphite, at least as far as the graphene plane is concerned, has also two atoms per unit cell but the symmetry is only hexagonal. Therefore, it has two optical modes, one nondegenerated Ag species and one twofold degenerated E2g species. The latter is located at §1600 cm1 and
94 is usually called the Gline. The other two lines, which we see in Fig. 3 for the graphitic material, are the Dline and the overtone of the Dline. The Dline is defectinduced and only observed if at least some disorder is present in the material.
Figure 3. Raman spectra for various carbon phases. Bottom: sp3 bonded diamond, second, third, and fourth from bottom: sp2bonded carbons with increasing disorder.
Figure 4. Raman line of the pentagonal pinch mode in C60 fullerene for various states of doping.
95 Interestingly, the overtone of this line is not defectinduced. Neither lines originate from *point phonons and are very famous in sp2 carbon systems. Figure 4 depicts Raman spectra of the pentagonal pinch mode of C60. In pristine material, this line is strongly resonantly enhanced, particularly in the greenblue spectral range, and located at 1468 cm1. If electron donors such as, e.g., alkali metals are intercalated, electrons are transferred from the metal to the fullerene orbitals and the conduction band fills up. This causes a quasilinear downshift of the pentagonal pinch mode by about 6.5 cm1 per transferred electron. This downshift has been used widely to determine the charge state of C60 in charge transfer C60 compounds. 2.3. SPECIAL TECHNIQUES IN RAMAN SCATTERING
2.3.1. Resonance excitation Since Raman scattering involves not only electronphonon interaction but also light induced electronic transitions, information on the latter can be obtained. This is in particular so for resonance scattering. A resonance profile of a Raman line represents the cross section as a function of excitation energy. Since all components of the Raman equipment are sensitive to the light energy, careful calibration is needed. This is usually done by measuring the spectrometer response of a wellknown cross section such as, e.g., the one of the 520 cm1 Raman line in Si. In this way, electronic transition energies can be determined even from powdered material or from thick, nontransparent films. 2.3.2. FanoBreitWigner line shape Another interesting possibility is the observation of Raman scattering from free electrons in metals or degenerate semiconductors. In this case, the HAAterm of the perturbation Hamiltonian becomes relevant and the excited Raman states exhibit a continuum. Unfortunately, the HAA perturbation is very small and the electronic Raman effect is therefore difficult to observe. However, if the electronic continuum couples to a phonon, interference effects deform the line shape of the latter and the response of the electron continuum becomes observable. The situation is depicted in Fig. 5. It shows the response from a Si crystal doped to various hole concentrations and excited with a red laser. For weak interaction (low doping), the Raman line becomes asymmetric and broadened. For stronger interaction, the line shape almost inverts.
96
Figure 5. FanoBreitWigner line shapes for the F2g mode in Si doped to various concentrations as indicated in units of 1020 cm3 (after Ref. [6]).
2.3.3. Dispersive modes According to the Raman concept, the distance of the Raman line from the exciting laser should be independent of the laser energy. This is not always the case. If the Raman shift depends on the energy of the exciting laser, we talk about dispersive Raman lines. There can be several reasons for a dispersion of the Raman lines as for example x
Photoselective resonance scattering in optically inhomogeneous material;
x
Carrier depletion at surfaces combined with light penetration effects;
x
Scattering from non *point phonons where the dispersion of the phonon mode becomes relevant and double resonance scattering applies.
In optically inhomogeneous material, the laser depicts predominately such parts where its energy matches to the optical gap. Since the vibrational frequencies often also scale with the optical inhomogeneity, dispersion is observed for such lines. Depletion of carriers happens at surfaces of semiconductors by surface state effects. This means that any contribution of the carriers to the phonon force
97 constants is modulated with the distance from the surface. Since the penetration of light depends on its energy, light of different energies will probe different regions of the crystal with respect to the distance from the surface and line dispersion will be observed. An example is shown in Fig.6 for pdoped Ge. Line dispersion for scattering from non *point phonons is very famous for the Dline in graphite and plays an important role for SWCNTs. Figure 7 depicts an experimental result for the shift of the Dline with laser excitation. The shift is approximately 43 cm1/eV, independent from the amount of disorder. The Dline represents a Kpoint phonon, which becomes Raman allowed by a doubleresonance scattering. The individual intermediate steps of the scattering processes are depicted in the insert of the figure. First, an electron–hole pair is excited under resonance conditions. Then a Kpoint phonon is generated and scatters the electron from K to K’, again in resonance, since it ends up at an eigenstate of the graphene bands. In order to satisfy qvector conservation, an electron is now elastically and nonresonantly backscattered to the Kpoint region by an impurity. From there it recombines nonresonantly with the hole. This doubleresonance scattering process is well known from semiconductor Raman scattering experiments7. In the case of overtone Raman scattering, which leads to the G'line in graphite, backscattering of the electron is established by another phonon. Therefore, no impurity scattering is required.
Figure 6. Raman line of pdoped Ge as excited by two different laser lines (after Ref. [8]).
98
Figure 7. Shift of the Dline in polycrystalline graphite versus laser excitation energy (after Ref. [9]). The insert depicts the graphite band structure close to the K and K'points of the Brillouin zone and the scattering process for the double resonance.
2.4. THE ELECTRONIC STATES OF SINGLE WALL CARBON NANOTUBES
2.4.1. Results from zone folding Zonefolding in the graphene plane is the simplest way to obtain the band structure of SWCNTs. This process will not be described here in detail, since other reports deal with it sufficient explicitly. General descriptions of the electronic properties of SWCNTs can be found in Refs. [1012]. Only properties with immediate connection to the Raman or resonance Raman process will be described. By the zonefolding process, a discrete set of curves is cut out from the band structure of graphene which represent the band structure of the SWCNTs. Maxima and minima in these H(k) relations lead to the set of van Hove singularities in the density of electronic states which determine, at least in a first approximation, the optical and resonance Raman behavior of the tubes. Figure 8 depicts the density of states for two selected tubes in a narrow energy range around the Fermi level. The (10,10) tube is clearly metallic and the (11,9) tube is semiconducting. In this approximation only transitions between symmetric Van Hove singularities are allowed and usually assigned as EiiS or EiiM for the semiconducting and metallic tubes, respectively. The energies EiiS and EiiM scale with the inverse tube diameter approximately as
99
Figure 8. Van Hove singularities for a (10,10) and for a (11,9) tube (left panel) and calculated joint density of states and Raman cross section for a (10,10) armchair tube (right panel).
Eii
2V0 a cc i D
S
0.85i , D
Eii
6V0 a cc i D
M
2.55i D
(6)
where V0 = 2.9 eV is the SSorbital overlap and aCC = 0.144 nm is the carboncarbon distance in graphene. There is a hierarchy if corrections to these values for the real tube states are accounted for. These corrections include: trigonal warping in the graphene band structure, chirality effects, curvature effects, and electron correlation (excitonic behavior). For the quantitative analysis of the tubes, these corrections are important as it will be discussed later. For the achiral tubes, the joint density of states can be evaluated analytically in the zone folding approximation. For armchair tubes (n,n) it yields
(H , n) g AC jds
¦ S aJ q
H 2 0
sin ª¬kq (H )a / 2(cos(qS / n) 2cos(kq (H )a / 2) º¼
(7)
100 where the sum runs over all subbands. This relation allows calculating the Raman scattering cross section explicitly using Eq. 6. The result of the integration is depicted in Fig. 8, right panel. 2.4.2. The Kataura plot The linear relation between the transition energies and the inverse tube diameter allows plotting the whole set of transition energies on a map called Kataura plot13 as shown in Fig. 9. The straight lines represent the relations of Eq. 5. The deviation of the points from the lines originates from trigonal warping and from chirality effects.
Figure 9. Linear relation between transition energies and inverse tube diameter. Each symbol represents one transition in one particular tube. Full circles are semiconducting tubes, open squares represent metallic tubes.
3. Raman scattering of singlewall carbon nanotubes Due to its strong resonance character, the Raman spectrum of SWCNTs is easily recorded and has very pronounced lines. The latter discriminates the SWCNT spectra from those of multiwall carbon nanotubes where usually only a broad Gline and a rather broad Dline, together with its overtone, is observed. For the SWCNTs each Raman line behaves in a special way as it will be
101 discussed below. The discussion will also be extended to filled tubes, especially for filling with fullerenes. 3.1. BASIC RAMAN LINES OF SINGLE WALL CARBON NANOTUBES
The Raman spectrum of SWCNTs has four characteristic lines: the radial breathing mode (RBM), the defect induced line (Dline), the graphitic line (Gline), and the overtone of the Dline (G'line). The D, G, and G'lines are correspond to lines in graphite but the RBM does not exist there (or has zero
Figure 10. Raman spectra of SWCNTs with a mean diameter of 1.36 nm as excited with 488 nm laser (upper spectrum) and 647 nm (lower spectrum).
frequency). Figure 10 depicts two Raman spectra of laser ablation grown SWCNTs as excited with two different lasers. The four characteristic modes are assigned. The inclined lines passing through the peaks of the D and G'lines indicate the dispersive behavior of these modes. The position of the RBM scales as 1/D where D is the tube diameter. This, together with the above described scaling of the transition energies, leads to a dispersive behaviour which manifests itself in a dependence of the line shape on the laser energy. The Gline has also some structure which will be discussed in detail below. Like in graphite, the intensity of the Dline increases with defect concentration whereas the G'line is independent of defects. The grouptheoretical approach is an important one for the Raman spectra analysis. In principle, line groups are needed to describe the symmetry properties of the tubes. The symmetry elements, point groups and number of
102 species of Raman active modes are summarized in Table 1. Each nanotube has in addition to the trivial translations a rotation axis and a screw axis parallel to the tube axis. The counting of the rotation axis is determined by the number of lattice points n(LP) on the chiral vector. This number is given by the greatest common divisor between m and n. In addition to the rotational axes along the tube axis, each tube has two types of twofold rotation axes perpendicular to the tube axis. Achiral tubes have some additional symmetry elements. Since we are mainly interested in q = 0 phonons, we can study the isogonal point groups related to the line groups which are Dn and Dnh for the chiral and for the achiral tubes, respectively, with n = N(R) = NR . NR is the number of hexagonal rings in the unit cell. In this way we find 8 and 14 Raman active modes for the achiral and chiral tubes, respectively. The Raman tensors are tabulated in classical textbooks like, e. g., Ref. [3]. The RBM is totallysymmetric and has A1g (A1) symmetry. The Gline has A1g, E1g, and E2g (A1, E1, and E2) components. Table 1. Summary of grouptheoretical properties of SWCNTs. The lower right graph depicts calculated Raman lines for a set of metallic SWCNTs (after Ref. [14]).
Symmetry elements: Line groups with SE T: translations, Cn(LP)s: rotations, 2S/n(LP), parallel Cn(LP)w: screw axes, S/n(LP) + a/2, parallel U, U': rotations, S, perpendicular Vh, Vv, Vh'g, Vh'rrfl: additional mirror planes, one mirror glide plane, and one rotoreflexion plan, only for achiral tubes
Point groups: For q = 0 modes, only the isogonal point groups to the line groups are relevant. Theses are DN(R) and DN(R)h for the chiral and achiral tubes, respectively, where N(R) is the number of rings in the unit cell Irreducible representations for Dn and Dnh: only 1dimensional and 2dimensional: A, B, E achiral tubes have center of inversion > g, u e.g. chiral tubes: A1, A2, B1 B2, E1, E2 ... EN(R)/2  1
Raman active modes: *ZRaman = 2A1g+ 3E1g+ 3E2g, zigzag *ARaman = 2A1g+ 2E1g+ 4E2g, armchair *CRaman = 3A1+ 5E1+ 6E2, chiral
The calculated frequencies in the lower right part of the table were obtained from zonefolding of the graphene plane (except for the RBM mode). As SWCNTs tend to assemble into bundles where tubetube interaction has some
103 influence on the frequencies, such frequencies may not immediately correspond to measured Raman modes. 3.2. SPECIAL RAMAN MODES
Now we will discuss the four special Raman modes in some detail. 3.2.1. The Radial breathing mode The RBM suffers to some extent from the tubetube interaction in the bundles. Therefore, its experimentally observed frequency is formally described by
Q RBM
C1 / D x C 2
(10)
where C1 and C2 are constants and x is very close to 1. Unfortunately, both constants vary from report to report in a rather wide range which makes it difficult to determine the tube diameter from the observed frequencies, even if Raman experiments were carried out on individual tubes. On the other hand, the Raman response from the RBM mode of a sample with distributed diameters exhibits an interesting oscillatory dispersion as depicted in Fig. 11, left panel. In the figure, the RBM Raman line patterns are plotted for a large number of exciting lasers. The origin of the oscillations is explained in the right panel of the figure. The straight lines on the right represent the Kataura plot. The dashed line on the left represents Eq. 8, without the constant C2. If a red laser excites a tube resonantly with transition H3, a RBM frequency will be observed as given by the dashed line connection. Shifting the laser energy upwards, the observed RBM frequency will also move upwards as thinner tubes will get into resonance. However, at some point the laser will also be able to excite in resonance rather large tubes with the next higher electronic transition, which would be H4 in the assignment of the figure. Then, the first moment of the total RBM response will start to decrease. Thus, the observed oscillations in frequency can be traced back to the jamming of the electronic states into van Hove singularities. The welldefined response of the RBM mode to the distribution of the tube diameters can be used to determine the diameter distribution function. In fact, by analyzing the first and the second moment of the RBM response and assuming a Gaussian distribution for the tube diameters, one can evaluate the mean and the variance of the Gaussian function.15 An even better and more reliable way is to measure the RBM response with two or three different lasers and obtain in a redundant manner the parameters for the Gaussian function from this procedure.
104
Figure 11. Raman response of the RBM as excited for many different lasers (after Ref. [15]). The wavy line is a guide for the eye (left); The right panel depicts a schematic for the resonance excitations. Eii ҏare the transition energies, r and b represent red and blue lasers. ȣ is the RBM frequency.
Recently some more detailed Raman experiments were carried out for the HiPco material. These tubes have usually a mean tube diameter of only 1 nm but a rather large width for the distribution function. Results for the RBM response are depicted in Fig. 12, left panel, as exited for various lasers. The oscillatory behavior is certainly also observed but a lot of individual lines appear, wide spread on the frequency axis. It is possible to measure the resonance profile for each of the peaks. The results can be plotted on a twodimensional grid of frequency and laser excitation energy. As a result we obtain the socalled “Raman chart”. An example for a Raman chart is depicted in Fig. 12. Another important result for the Raman response comes from spectra recorded for individual tubes. Since the resonance crosssection was large, the spectra could be measured from a scattering volume as low as 106 Pm3. This is the volume of one SWCNT in a laser focus of 500 nm diameter. Figure 13 gives an example. The line from Si originates from an impurityactivated mode and is
105 about 50 times smaller than the Raman response of the Si optical mode. Estimating the scattering volume for Si to be 0.2 Pm3 we find the Raman cross section of SWCNTs more than a factor 106 larger as compared to Si.
Figure 12. Raman response for the RBM mode of HiPco tubes (left) (after Ref. [16]). Raman cross sections as plotted over a 2D grid of RBM frequency and laser excitation energy (after Ref. [17]).
Figure 13. Raman spectra of the RBM modes at 148, 164, and 237 cm1, respectively. The line at 300 cm1 comes from Si (after Ref. [18]).
106 3.2.2. The Gline From a theoretical consideration, the Gline in SWCNTs consists of six components with A1, E1, and E2 symmetry and longitudinal or transversal (= circumferential) orientation of the normal coordinates. For the achiral tubes, the index g must be added as usual. Experimentally indeed a more or less expressed fine structure has been reported for the Gline pattern. An exact assignment of the fine structure to the A and E modes has not been possible so far. The Gline does not exhibit dispersion except when it comes to the excitation of metallic tubes. However, the Gline exhibits an oscillatory behavior with respect to line intensity if the laser energy is tuned. This is demonstrated Fig. 14, left panel. The set of spectra represent the Gmode response as excited for various lasers. Starting with the highest laser energies, the line intensities increase down to about 2.49 eV. With the further reduction of the energy, the main peak of the Gline decreases but a broad structure appears about 40 cm1 downshifted.
Fiure 14. Left: Raman response for the Gline as excited for various lasers. The spectra are normalized in intensity (after Ref. [19]); Right: Kataura plot of transition energy versus RBM frequency, including curvature effects (after Ref. [20]). The horizontal lines are energy markers.
This broad feature peaks for about 1.93 eV excitation and decreases again for further reduction of the excitation energy. Simultaneously, the main peak of the Gline starts to grow again. The right part of the figure depicts a modified
107 Kataura plot. The calculation was performed on the level of symmetryadapted nonorthogonal tightbinding model, which allows one to consider curvature effects. The deviations from the linear behavior are particularly strong for the small diameter tubes (high frequency RBMs). For the tubes used in Fig. 14, left, the mean diameter was 1.4 nm, which corresponds to an approximate RBM frequency of 180 cm1. The three horizontal lines mark the energies where the main Gline peaks for the first time, where the broad sideband peaks, and where the main Gline peak approaches a peak again. Considering the diameter of the tubes (or the mean RBM frequency of 180 cm1) it is obvious, that the first peak of the Gline coincides with the resonance at the E33S transition. The maximum of the broad sideband occurs for resonance excitation of metallic tubes at the E11M transition and the reincrease of the main peak takes place when the laser energy comes down to the E22S transition for semiconducting tubes. A blowup of the response from the metallic tubes exhibits a FanoBreitWigner line shape demonstrating the strong coupling of this particular line with the free carriers in the metallic tubes.
Figure 15. Frequencies for the Gline as calculated using the Vienna Abinitio Simulation Package (VASP) (after Ref. [21]). The full drawn line is an empirical approximation.
The frequencies of the Gline components were calculated for armchair and zigzag tubes using the densityfunctional theory code VASP (Vienna Abinitio Simulation Package). The results are depicted in Fig. 15 for the three selected components A1(T), A1(L), and E2(L). Experimental results for the line position of the metallic tubes are also included in the figure. Two results are striking.
108 The dependence of line frequency on tube diameter is small for the A modes as long as the tube diameters are rather large. Only for very small tubes, a considerable deviation from the limiting value for large tubes is observed. Only the E2(L) mode exhibits a considerable down shift with decreasing tube diameter. The other important result is the dramatic drop of line frequencies for the metallic species of the A1(L) modes. This mode softening corresponds exactly to the experimentally observed downshift of the Raman response for the metallic tubes. A special Peierlstype electronphonon coupling mechanism was found to be responsible for this softening. This mechanism operates only for the A1(L) frequencies. The full drawn line represents an empirical law of the form Q = 1619 – 866/D (Ref. [22]). There were several reports concerning the change of the Gline Raman response on doping with electron donors and electron acceptors. The general observation is a quenching of the resonance enhancement as a consequence of filling the conduction band (or depleting the valence band). Figure 16 depicts an example where arcdischarge grown SWCNTs were intercalated with Li. A dramatic decrease of the lower frequency components and a weak upshift of the high frequency component are observed. Obviously, the metallic tubes run out of resonance more rapidly.
Figure 16. Raman response of the Gline as excited with two different lasers after doping with Li. The Li concentration increases from top to bottom (after Ref. [23]).
109 3.2.3. The Dline and the G'line The Dline and the G'line exhibit a quasilinear linear dispersion. A set of spectra for both lines is depicted in Fig. 17. The intensity profiles represent again the resonance transitions from the Kataura plot. The right part of the figure shows explicitly the shift of the lines with increasing laser energy. For comparison, the completely flat dispersion for the Gmode is also depicted. The slope of the quasilinear behavior is 42.3 cm1/eV for the Dline and 86.4 cm1 /eV for the G'line. Both values are very similar to the values for graphite. Therefore, it is obvious that the dispersion for the SWCNTs originates from Kpoint phonons like in graphite. However, since there is a difference in the electronic structure between the two systems, one would expect also some difference in the dispersion. This is indeed the case as we can observe an oscillatory behavior on top of the linear relation. A quantitative analysis revealed indeed that the oscillations originate from the jamming of the electronic states into van Hove singularities.24 This means that the oscillatory behavior is a consequence of the triple resonance effect relevant for the SWCNTs.
Figure 17. Raman spectra for the Dline (left) and for the G'line (center) as excited for various lasers. The right part of the figure depicts peak positions of the lines versus laser energy. equivalent to G’. The Gline is included for comparison; after 24 .
110 3.3. FILLING CARBON NANOTUBES (PEAPODS)
One of the many interesting features observed for SWCNTs is the possibility to fill the tubes with other materials. Many different fillers were successfully incorporated into the tubes so far. One of the most exciting filling materials is fullerenes. There is obviously a very strong cohesive force which attracts the fullerenes to the interior of the tubes. The classical and historically first observation of fullerenes inside the tubes came from transmission electron microscopy (TEM).25 However, since the tubes are at least partly transparent to visible light, a Raman response from the fullerenes inside the tubes can also be expected. This has been demonstrated in various reports even though some of the fullerene modes appear as very weak lines. The line position of the pentagonal pinch mode is downshifted by 3 wave numbers as compared to the response from the free C60 molecule and exhibits a characteristic line splitting.26
Figure 18. Raman spectra of doped peapods in comparison to undoped peapods and doped pristine tubes; Gline for doped peapods as compared to doped pristine tubes for three different laser excitations (left); RBM spectral region for doped peapods as compared to polymeric fullerenes and Rb6C60 (center); blow up of RBM spectral region for doped peapods (right) (after Ref. [27]).
Like the empty tubes, the peapods can be doped with electron donors such as Rb. For a reasonably strong doping not only the Raman response from the tubes is modulated as expected from Fig. 16 but also the response from the pentagonal pinch mode is downshifted. For heavy doping the mode shifts by 37 cm1 which corresponds roughly to a charge transfer of 6 electrons. Thus, C60 has adopted a charge state –6. This is demonstrated in Fig. 18, left, for excitation with three different lasers. The simultaneously doped pristine tubes do not exhibit the fullerene line. The center of the figure depicts the radial part
111 of the spectrum and compares the response from the doped peapods with the Raman spectrum of Rb6C60 and the orthorhombic polymeric phase RbC60. In the peapod spectrum two additional lines appear which are assigned as P and clearly indicate that the C606 molecules inside the tubes are polymerized. Calculations on the tightbinding level suggested that this polymer is single bonded.28 Interestingly, due to the doping, the resonance behavior for the RBM of the tubes is lost and with it also the dispersion. This is demonstrated in the right panel of Fig. 18. 4. Raman scattering of doublewall carbon nanotubes Doublewall carbon nanotubes (DWCNTs) are another new species in the family of carbon nanophases. They are of particular interest if grown from peapods, since then there is exactly one inner tube in one outer tube. DWCNTs exhibit several remarkable features. They have the same small size as the SWCNTs but are stiffer and therefore more appropriate as mechanical sensors. Field emission is enhanced. The inner tubes have a very high curvature and allow therefore studying curvature effects in great detail. Inner shell tubes are grown in a highly shielded environment without catalyst and are therefore highly unperturbed. The outer tubes may be subjected to functionalization while the inner tubes remain untouched. Alternatively, as it will be shown below, the inner tubes may be modified, e.g., by isotope substitution whereas the outer tubes remain as grown. 4.1. THE RAMAN RESPONSE FROM THE INNER TUBES
The growth of inner tubes from peapods by annealing at high temperatures was first demonstrated by Smith et al.29 and proven by TEM. Like in the case of the peapods, Raman spectroscopy is another possibility to study the transition process. Figure 19 depicts a series of spectra which demonstrate the transition from peapods to DWCNTs. Annealing was performed for two hours in high vacuum at 1500 K. The difference in the spectra recorded for peapods and DWCNTs clearly demonstrate the transition. In the spectrum of the DWCNTs, all lines from the fullerenes have disappeared and a new set of very sharp lines have appeared at around 300 cm1. As the inner tubes have a much smaller diameter, this set of lines is obviously the response from the RBM of the latter. Note that for the red laser, the response from the inner tubes is much larger than the response from the outer tubes even though there is much less carbon material in the former. The right part of the figure depicts a histogram of the diameter distribution. It was drawn under the assumption that the wall to wall distance between inner and outer tube is 0.36 nm which is very close to the
112 plane to plane distance in graphite. From this figure, three important results can be deduced: Firstly, the number of carbon atoms in the inner tubes is only about one third of the carbon atoms in the outer tubes. Secondly, the distribution of the geometrically allowed tubes is much coarser for the inner tubes as compared to the outer tubes. And finally, there is a cutoff for the inner tube diameters just below d = 0.5 nm. The mean diameter of the inner tubes is about 0.7 nm.
Figure 19. Left: Raman spectra for SWCNTs, as exited with 514 nm (bottom), for peapods as exited with 488 nm (center) and for DWCNTs as excited with 647 nm (top). Right: Diameter distribution plotted as two histograms for outer shell and inner shell tubes and as a curve for HiPco material.
The sharp lines recorded for the Raman response from the inner tubes are surprising. The measurements of the response from these tubes at low temperature and in the high resolution mode of the spectrometer revealed line widths down to 0.4 cm1 after deconvolution from the spectrometer response. An example is depicted in Fig. 22. The narrow lines indicate a long phonon life time and therefore a highly unperturbed tube material. In addition to the small line width, the figure also demonstrates a characteristic line splitting. As a result of this splitting, the number of observed lines is considerably larger than the number of geometrically allowed tubes which can be accommodated in the interior of the host tubes. To a first approximation, this can be explained by the lower density of the geometrically allowed tubes with smaller diameter. Several outer tubes with increasing diameter may be considered as a host before the next larger inner tube can grow for best matching conditions. Since the tubetube interaction shifts the RBM frequency and since this interaction depends on the wall to wall distance between the tubes, some splitting of the lines can be expected. (See also the discussion below.)
113
Figure 20. Raman response of the RBM from the inner tubes as recorded with a resolution of 0.5 cm1 (top) and after deconvolution from the spectrometer response (bottom). The arrows indicate line splitting.
4.2. MODE ASSIGNMENT AND HIGH CURVATURE EFFECTS
Recording the Raman response with a large number of laser lines revealed the Raman response from all inner tubes in the diameter range from 1.1 nm down to 0.5 nm. Since the distribution of the RBM frequencies is not a smooth function of the diameters, a best fit search for the assignment of the lines to the (n,m) values of the corresponding tubes was possible. The result was reported in Ref. [30] and revealed a welldefined straight line for the relation between RBM frequency and inverse tube diameter of the form QRBM = C1/D + C2 with C1 = 234 cm1nm and C2 = 14 cm1. A more dramatic demonstration of the high curvature effects can be obtained from an analysis of the high frequency modes such as the G' mode. The Raman response for this mode as excited with various lasers is depicted in Fig. 21. The dispersion effect is clearly visible for the outer shell tubes (higher frequency peak) and for the inner shell tubes (lower frequency peak). A
114 quantitative display of the dispersion is depicted in the right panel of the figure. The slopes for the inner and outer tubes are 85.4 and 99.1 cm1/eV, respectively. From the analysis of the tubes with different diameters in the range of 1 to 1.4 nm, the empirical relation of the form
Q G'
2536 21 / D
(5)
was obtained in Ref. [31] for the dependence of the G' frequency on the tube diameter. This relation yields the correct value of 2684 cm1 for the outer tube if the spectrum is excited with a green laser but a much too high value of 2658 cm1 for the inner tubes. The observed value is 2628 cm1 which means 30 cm1 lower as expected from the extrapolation of Eq. 5. This difference is considered as a breakdown of Eq. 5 for high curvature tubes.
Figure 21. Left: Raman response of the G' mode of DWCNTs for various lasers as indicated. The high and the low frequency peaks come from the outer and inner shell tubes, respectively. The right panel of the figure depicts the dispersion of the two modes. The value of 2658 cm1 for excitation with a green laser (2.42 eV) was obtained for an extrapolation from low curvature tubes.
4.3. ISOTOPE SUBSTITUTION
It was demonstrated recently that the inner tube can be grown as in a highly 13C enriched form by using 13C enriched fullerenes as a filling material32. From the mass enhancement all vibrational modes are expected to shift downwards according to the concentration of 13C used. This is demonstrated in Fig. 22, left, for the RBM. The downshift demonstrates the mass enhancement. The 13C enhanced inner shell tubes exhibit dramatic advantages for 13C NMR spectroscopy. Not only that the signal is enhanced by a factor of about ten but it
115 is also highly selective and originates almost only from the latter. Any response from contaminating carbon phases is strongly suppressed. Figure 24, right, demonstrates this. The top spectrum in the figure displays the response for unsubstituted SWCNTs together with the expected line shape resulting from the chemical shift anisotropy of the NMR gfactor. The two lower spectra represent the static and the magic angle spinning response for the highly 13C enriched inner tubes. Both lines are downshifted to about 111 ppm as compared to 124 ppm for the unsubstituted tube and are also broadened. Both effects are at least partly due to high curvature. The downshift reflects the approach to a sp3 hybridization for which the 13C NMR resonance is known to appear at 65 ppm. A contribution of the shielding of the magnetic field due to the outer tubes can also contribute to the downshift. The line shape for the response from the static NMR does not represent any more the chemical shift anisotropy. It is rather distorted from the distribution of small tube diameters which correlates to a distribution of sp2 admixture. The same argument holds for the line broadening of the MAS response.
Figure 22. 13C substitution in carbon nanotubes; Left: Raman response for the RBM line; Right: NMR spectroscopy of carbon nanotubes. Top: Response of static NMR for non enriched SWCNTs. The full drawn line is a calculated fit. Center and bottom: static NMR and MAS NMR, respectively, for highly 13C enriched inner shell tubes.
116 4.4. CLUSTERING OF MODES AND PAIR SPECTRA
The Raman response for the RBM from the inner tubes revealed too many lines which could partly be explained by the different density in the geometrically allowed inner and outer tubes. To take care of the large number of tubes, however, also combinations of inner – outer pairs with not optimized wall to wall distance must be considered. In addition, a comparison of the inner tube RBM spectra to the RBM response from HiPco tubes in the same diameter range revealed a significant absence of lines in the latter material as compared to the signals from the inner shell tubes. This is demonstrated in Fig. 23. In addition, the values obtained for C1 and C2 from the assignment procedure described above exhibited some discrepancies with a recent analysis of the RBM in HiPco tubes.17,33 It was therefore desirable to get a complete frequency and crosssection analysis for the RBM line of the inner tubes. Due to the small diameter of the inner tubes, the tightbinding derived relation between RBM frequency and optical transition energies (Kataura plot from Fig. 9) was expected to be insufficient for this purpose, since it does not include curvature effects. Instead, it turned out that extended tight binding calculations using symmetry adapted non orthogonal tight binding wave functions as reported by V. Popov20 is a better approach. It reveals dramatic differences to the simple tight binding calculation for tube diameters smaller than 1 nm. Results from such calculations are depicted in Fig. 24 in comparison to results from HiPco tubes. Several striking information can be obtained.
Figure 23. Raman response for the RBM of the inner tubes (upper spectrum) as compared to the response from HiPco tubes (lower spectrum).
117
Figure 24. PopovKataura plot for the relation between transition energy and RBM frequencies as obtained from an extended tightbinding calculation and HiPco material. Open symbols: averaged for HiPco tubes17,33 or calculated values. Lines: connect tubes from the same family with family number indicated. Full symbols: observed values for the peak resonance of the inner tubes.
1. The response from the inner tubes follows the same family behavior as the response from the HiPco tubes but the transition energy is generally downshifted by about 50 meV. 2. There is a considerable number of lines which exhibit almost the same resonance transition energy even though the frequencies spread out as much as 30 cm1. 3. The number of lines in the cluster and thus the width of the cluster increases with decreasing tube diameter. 4. The low frequency end of the clustered lines matches well to the results from the calculation and to the experimental results from HiPco tubes. Therefore the cluster can be assigned to one particular inner tube which has the right resonance energy. This effect is best seen for tube (6,4) of family 16 and tube (6,5) of family 17. These tubes have 14 members in the cluster
118 as determined from a high resolution measurement at low temperatures (see also Fig. 25). 5. Within the cluster, the resonance energies are slightly decreasing with increasing RBM frequency. The averaged value for the decrease is –2 meV/cm1 as indicated by the dashed line in the figure.
Figure 25. Upper panel: RBM frequencies (xaxis) versus difference in tube diameters (yaxis) as calculated for a (6,4) inner tube and various outer tubes. Lower panel: composed spectrum as observed for the (6,4) cluster (full line) and calculated spectrum where the frequencies from the upper part of the figure were dressed with Lorentzian lines and the intensities were weighted by a Gaussian distribution.
The results can be interpreted in a sense that one inner tube can grow in many, up to 14, different outer tubes and the radial tubetube interaction is responsible for the line shift. In other words, the outer tubes exert pressure on the inner tubes which causes the line shift. The line shift can be calculated in a continuum model approximation where the tubetube interactions were described by a LennardJones potential and the dynamical matrix for the frequencies was solved explicitly under these conditions. Results of the calculation are depicted in the upper part of Fig. 25. The lower part of the figure shows the spectrum composed of the contributions of the inner tubes dressed by a Voigtian line. In this way, the spectrum directly represents the population of the pairs of the inner (6,4) tube and the various outer tubes in the cluster. The dotted spectrum was obtained by using the calculated frequencies and dressing them with
119 Lorentzian lines. The intensities of the lines were taken from an assumed Gaussian distribution. In detail, the logic is a bit more complicated. During the growth process, various inner tubes are able to grow in a given type of outer tube. This holds for a reasonably large number of outer tubes. Thus, the (6,4) tube exists in a large number of different outer tubes but there is also a large number (maybe ten times more) of the same outer tubes which do not host a (6,4) tube. It is the process of photoselective Raman scattering which singles out the various (6,4) pairs. In other words, the very strong Raman response of the (6,4) tubes comes only from approximately one tenth of the tube pairs. The higher electronphonon coupling and the longer lifetime of the excited state are probably the reason for this resonance Raman effect. 5. Summary In summary, we presented in this contribution some fundamental aspects of Raman scattering of carbon nanotubes with particular emphasis on resonance excitation. A classical and a quantummechanical description are provided. The dispersion of the Raman lines and FanoBreitWigner interferences are discussed. In the second part of the manuscript, an extensive description of the Raman scattering from SWCNTs is given. The origin of the quantum oscillation for the RBM, the amount of crosssectional enhancement and the diameter dependence of the graphitic mode are discussed. This part also contains a review on the Raman response of doped nanotubes upon and of peapod systems. Finally, in the last chapter, the special topic of Raman scattering from the inner shell tubes in DWCNTs is summarized. These systems are special since the inner tubes have a high curvature and an extremely strong Raman effect. By substitution of 12C of the inner tubes with 13C carbons, challenging effects in Raman line shifts and in NMR spectroscopy are reported. ACKNOWLEDGEMENTS
Work supported by the FWF Nr. 17345 and EU BIN2200100580 and MEIFCT2003501099 projects. Valuable discussions with V. N. Popov are gratefully acknowledged.
References 1.
M. Cardona, Light Scattering in Solids, Topics in Applied Physics Vol 8, (Springer, Berlin, Heidelberg, 1983)
120 2.
H. Poulet , J.P. Mathieu, Vibrational Spectra and Symmetry of Crystals, (Gordon &
Breach, Paris, 1970) 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.
H. Kuzmany, Solid State Spectroscopy, (Springer, Berlin, Heidelberg 1999) G. Turell, Infrared and Raman spectra of Crystals, (Academic, London, 1972). D.A. Long, Raman Spectroscopy, (McGrawHill, New York 1977). G. Abstreiter, in Light Scattering in Solids IV, M. Cardona, G. Güntherrodt (eds.) Topics in Appl. Phys. Vol 54 (Springer, Berlin, Heidelberg 1984). M.L. Bansal, A.K. Sood, and M. Cardona, Solid State Commun. 78, 579582 (1991). D. Olego and M. Cardona, Phys. Rev. B 23, 65926602 (1981). C. Thomsen and S. Reich, Phys. Rev. Lett. 85, 5214 (2000). M. S. Dresselhaus, G. Dresselhaus, and Ph. Avouris, Carbon Nanotubes: Synthesis, Structure, Properties and Applications (SpringerVerlag, Berlin, 2001). R. Saito, G. Dresselhaus, and M. S. Dresselhaus, Physical Properties of Carbon Nanotubes, Imperial College Press (London), 1998. S. Reich, C. Thomsen, J. Maultsch, Carbon Nanotubes, (WileyVCH, Weinheim, 2004). H. Kataura, Y. Kumazawa, Y. Maniwa, I. Umezu, S. Suzuki, Y. Ohtsuka, and Y. Achiba, Synth. Met. 103, 25552558 (1999). R. A. Jishi, L. Venkataraman, M. S. Dresselhaus, and G. Dresselhaus, Phys. Rev. B 51, 1117611179 (1995). H. Kuzmany, W. Plank, M. Hulman, Ch. Kramberger, A. Grüneis, Th. Pichler, H. Peterlik, H. Kataura and Y. Achiba, Eur. Phys. J. B 22, 307320 (2001). A. Kukovecz, Ch. Kramberger, V. Georgakilas, M. Prato and H. Kuzmany, Eur. Phys. J. B 28, 223230 (2002). C. Fantini, A. Jorio, M. Souza, M. S. Strano, M. S. Dresselhaus, and M. A. Pimenta, Phys. Rev. Lett. 93, 147406/14 (2004). A. Jorio, R. Saito, J. H. Hafner, C. M. Lieber, M. Hunter, T. McClure, G. Dresselhaus, and M. S. Dresselhaus, Phys. Rev. Lett. 86, 1118/14 (2001) A. Grüneis Diploma work , University of Vienna 2001. V. N. Popov, New J. Phys. 6, 17/117 (2004). O. Dubay, G. Kresse, H. Kuzmany, Phys. Rev. Lett. 88, 235506/14 (2002). R. Pfeiffer, 2004, unpublished. J. L. Sauvajol, 2004, unpublished H. Kuzmany, W. Plank, and M. Hulman, Adv. Solid State Phys. 40, 194 (2000). B. W. Smith, M. Monthioux, and D. E. Luzzi, Nature 396, 323329 (1998). R. Pfeiffer, H. Kuzmany, T. Pichler, H. Kataura, Y. Achiba, M. MelleFranco, and F. Zerbetto, Phys. Rev. B 69, 035404/17 (2004) T. Pichler, H. Kuzmany, H. Kataura, and Y. Achiba, Phys. Rev. Lett. 87, 267401/14 (2001) S. Pekker, G. Oszlanyi, G. Faigel, Chem. Phys. Lett. 282, 435438 (1998). B. W. Smith, and D. E. Luzzi, Chem. Phys. Lett. 321 169172 (1999). Ch. Kramberger, R. Pfeiffer, H. Kuzmany, V. Zólyomi, and J. Kürti, Phys. Rev. B 68, 235404/14 (2003). R. Pfeiffer, H. Kuzmany, F. Simon, S. N. Bokova, and E. Obraztsova, Phys. Rev. B 71, 155409/18 (2005). F. Simon, Ch. Kramberger, R. Pfeiffer, H. Kuzmany, V. Zólyomi, J. Kürti, P. M. Singer, and H. Alloul, Phys. Rev. Lett. 95, 017401/14 (2005). H. Telg, J. Maultzsch, S. Reich, F. Hennrich, C. Thomsen, Phys. Rev. Lett. 93, 177401/14 (2004).
RAMAN SPECTROSCOPY OF ISOLATED SINGLEWALLED CARBON NANOTUBES THIERRY MICHEL,* MATTHIEU PAILLET, PHILIPPE PONCHARAL, AHMED ZAHAB, JEANLOUIS SAUVAJOL UMR CNRS 5587, Université Montpellier II, Montpellier, France JANNIK C. MEYER, SIEGMAR ROTH Max Planck Institute for Solid State Research, Stuttgart, Germany
Abstract. Raman spectra of two isolated singlewalled carbon nanotubes with close diameters but different chiral angles are presented and discussed. Keywords: isolated singlewalled carbon nanotubes; Raman spectroscopy
The resonant Ramanscattering technique is a powerful tool for investigating singlewall carbon nanotubes (SWCNTs), for which the radial breathing mode (RBM) and the tangential modes (TM) are the two main features. Recently, several experiments have been devoted to the study of the Raman spectrum at the singlenanotube level (Reich et al. 2004). Here we present a complete experimental procedure and results obtained on two isolated individual SWCNTs with close diameters but different chirality angles. We developed a complete procedure including the preparation of the substrates, the sample preparation, AFM imaging and Raman spectroscopy (Paillet et al., 2005). Isolated SWCNTs were grown by chemical vapor deposition (Paillet et al. 2004) on lithographically marked substrates. Atomic force microscopy (AFM) was used to check the SWCNT density (< 1 per µm²), their positions and orientations with respect to the micrometric markers (Fig. 1, left). Roomtemperature microRaman experiments were performed using the Ar/Kr laser lines at 514.5 nm (2.41 eV) and 647.1 nm (1.92 eV) on a JobinYvon T64000 spectrometer. The Raman spectra of two isolated SWNTs of the same diameter (RBM at 186.4 and 187 cm1), acquired with two different laser wavelengths (2.41 and 1.92 eV), indicating that one is semiconducting and the other is metallic (Reich et al., 2004), are presented in Fig. 1, right. The profile of the TM of the metallic tube is closer to that of the semiconducting tube than *To whom correspondence should be addressed. Thierry Michel; email: [email protected]
121 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 121–122. © 2006 Springer. Printed in the Netherlands.
122 reported for SWCNT bundles (Reich et al., 2004). Slight differences remain however in the peak positions, relative intensities and widths. These results are a strong experimental evidence for the chirality dependence of TM lineshape and the vanishing of the BreitWignerFano component in isolated metallic SWCNTs (Paillet et al., 2005).
800nm
Figure 1. Left: typical AFM image of a low density SWCNT sample synthesized by CVD. Right: Raman spectra of two individual SWCNTs collected using laser photon energy of 1.92 eV (top curves) and 2.41 eV (bottom curves).
In summary, we performed a Raman study of isolated SWCNTs showing the tricky influence of their atomic structure. Experiments involving the combined use of electron diffraction and Raman spectroscopy on the same isolated SWNT are in progress (Meyer et al., 2005) and should allow determining unambiguously the intrinsic response of a specific SWCNT.
References Reich, S., Thomsen, C., Maultzsch, J., 2004, Carbon Nanotubes: Basic Concepts and Physical Properties, John Wiley and Sons, 320 pp. Paillet, M., Poncharal, P., Zahab, A., Sauvajol, J.L., Meyer, J. C. and Roth, S., 2005, Vanishing of the BreitWignerFano component in individual singlewall carbon nanotubes, Phys. Rev. Lett. accepted. Paillet, M., Jourdain, V., Poncharal, P., Zahab, A., Sauvajol, J.L., Meyer, J. C., Roth, S., Cordente, N., Amiens, C. and Chaudret, B. 2004, Versatile synthesis of individual individual singlewalled carbon nanotubes from nickel nanoparticles for the study of their physical properties, J. Phys. Chem. B 108:1711217118. Meyer, J. C., Paillet, M., Sauvajol, J.L., Duesberg, G. S. and Roth, S. 2005, Lattice structure and vibrational properties of the same nanoobject, http://arxiv.org/condmat/0501341.
Part III. Electronic and optical properties and electrical transport
ELECTRONIC TRANSPORT IN NANOTUBES AND THROUGH JUNCTIONS OF NANOTUBES PH. LAMBIN* Département de Physique, Facultés Universitaires NotreDame de la Paix, 61 Rue de Bruxelles, B 5000 Namur, Belgium F. TRIOZON DRT/LETI/DIHS/LMNO, CEA, 17 Rue des Martyrs, F 38054 Grenoble Cedex 9, France V. MEUNIER Computer Science and Mathematics Division, Oak Ridge National Laboratory, Oak Ridge, TN 378316367, USA
Abstract. This chapter briefly reviews the electronic properties of singlewall carbon nanotubes as described by the ʌelectron tightbinding approximation. The Landauer formalism is introduced next and applied to the study of quantum transport in carbon nanotubes. The electronic properties and electric conductance of intramolecular nanotube junctions are also discussed.
Keywords: electronic structure; density of states; conductivity; transport properties; intramolecular junctions
1. Introduction The pace of research on carbon nanotubes has been considerably enhanced by the discovery of the extraordinary electronic properties predicted for singlewall nanotubes (SWNTs). These properties are very sensitive to the nanotube atomic structure, namely its diameter and chiral angle (Mintmire et al., 1992; Hamada et al., 1992), or conventionally to the wrapping indices n and m that label the arrangement of the honeycomb atomic lattice on a seamless cylinder. It was readily predicted that when the difference n – m between the wrapping indices is a multiple of 3, the nanotube is metallic, otherwise there is a semiconducting gap in its band structure (Hamada et al., 1992; Saito et al., 1992). These *To whom correspondence should be addressed. Philippe Lambin; email: [email protected]
123 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 123–142. © 2006 Springer. Printed in the Netherlands.
124 properties, and other, can be deduced easily by a simple zone folding of the ʌbands of graphite, as demonstrated in Sec. 2. The extreme sensitivity of the electronic properties to the atomic structure of the nanotube is a consequence of the quantization of the ʌelectron wavevector imposed to the wavefunction by the cyclic boundary condition around the circumference. It follows that a slight change of the couple (n,m) can result in the SWNT being either metallic or semiconducting. Possible applications of nanotubes in nanoelectronic devices will have to accommodate and exploit that property (Avouris, 2002). On a more fundamental point of view, the fact that the nanotubes are onedimensional systems brings about interesting phenomena. More precisely, carbon nanotubes are quasi onedimensional systems: the atoms are located on a cylindrical surface in three dimensional space; they do not form a onedimensional chain, like for instance in polyacetylene. This is important since the Peierls instability responsible for alternating single and double bonds in polyacetylene for example is generally inoperant in the case of carbon nanotubes, except for very small diameter (Connétable et al., 2005). Other onedimensional effects are predicted to occur in metallic nanotubes, such as the breakdown of the usual Fermi liquid description in favor of TomonagaLuttinger behavior dominated by collective excitations (Bockrath et al., 1999). Nanotubes are remarkable ballistic conductors (Frank et al., 1998). This again is a consequence of the atoms being distributed around the circumference of the tube, which averages defectinduced scattering to a small value. The charge transport in these quasi onedimensional systems is of quantum origin. It can be described by the Landauer theory, whose formalism is reviewed and illustrated in Sect. 3. Seamless connections between singlewalled nanotubes with different helicities can be realized easily, at least on the computer. These socalled intramolecular junctions have interesting electronic properties, which are briefly described in Sect. 4. 2. Zonefolding approximation of graphene band structure Conceptually, a singlewall nanotube is a rolledup strip of graphene of which it retains most of its electronic structure. Close to the Fermi energy, the band structure of graphene is dominated by the ʌ states formed by the interacting 2pz orbitals normal to the graphene sheet (Wallace, 1947). In the Linear Combination of Atomic Orbital (LCAO) approximation, the wavefunction is written as a linear combination of 2pz atomic orbitals ĭJ located on atoms J of the graphene sheet, ȥ = ȈJ CJĭJ. Assuming that the overlap between neighboring orbitals can be neglected, the Schrödinger equation Hȥ = Eȥ is thus equivalent to a set of linear equations for the LCAO coefficients CJ
125
¦)
I
H ) J CJ
ECI
(1)
J
on any site I. Due to the periodicity of the graphene layer, the Bloch theorem can be used to relate any coefficient to those of the two atoms in the unit cell shown in Fig. 1(a), CA and CB, by simply multiplying the latter by the Bloch factor exp(ik·T) where T is the appropriate translation vector. Equation 1 therefore becomes a set of two linear equations for CA and CB. Restricting further the hopping interactions ) I H ) J to first neighbors yields
H S C A J 0 F * (k )CB J 0 F * (k )C A H S CB
EC A
(2a)
ECB
(2b)
)A H )A )B H )B where H S is the onsite energy J 0 ) A H ) B is the hopping interaction. In these equations, F (k ) 1 exp(ik a1 ) exp(ik a 2 )
and (3)
with a1 and a2 two primitive translation vectors of the graphene hexagonal lattice (see Fig. 1(a)). The two eigenvalues of the set of equations (Eq. 2), given by Er
H S r J 0 F (k ) ,
(4)
correspond to the bonding ʌ (minus sign) and the antibonding ʌ (plus sign) bands.
Figure 1. (a) Graphene plane with two primitive translation vectors a1 and a2 and its two atoms A and B in a unit cell. (b) FirstBrillouin zone (FBZ) of graphene with its three highsymmetry points ī, M, and K. The shaded area corresponds to the irreducible part of the FBZ.
126 The separation between the ʌ* and ʌ bands in the reciprocal plane of graphene is shown in Fig. 2(b). These bands cross each other at the corners of the hexagonal firstBrillouin zone shown in Fig. 1(b), conventionally labeled K. Indeed, setting kK
1 2 b1 b 2 3 3
(5)
with b1 and b2 the reciprocal unit vectors, leads to F (k K ) 1 exp(i 2S / 3) exp(i 4S / 3) , which is zero as the sum of the three complex cubic roots of unity. At the corners of the firstBrillouin zone, E E H S . For symmetry reasons, the Fermi level coincides with the energy of the crossing. Graphene is a zerogap semiconductor; its density of states (DOS) at the Fermi energy EF is zero and increases linearly on both sides of EF. In a singlewall nanotube, cyclic boundary conditions apply around the circumference. In the planar development shown in Fig. 2(a), the circumference of the nanotube is a translation vector of graphene, C na1 ma 2 , n and m being the wrapping indices. In the twodimensional graphene sheet, the wave function at the end point of C is the one at the origin multiplied by the Bloch factor exp(ik·C). Assuming that the graphene wavefunctions is unaltered in the rolled up structure (zone folding approximation), the cyclic boundary conditions impose exp(ik·C) = 1. That condition discretizes the Bloch vector k along equidistant lines C·k = 2ʌl perpendicular to C and therefore parallel to the nanotube axis, where l is an integer number. These discretization lines are drawn in Fig. 2(b), for the (5,3) nanotube. Along each of these lines, ʌ* and ʌ bands of graphene are sampled. Both bands are always separated by a gap, unless one of the discretization lines passes exactly through a corner K of the firstBrillouin zone. The nanotube is then a metal, because it contains two bands that cross the Fermi level. The condition for that is C·kK = 2ʌl. With the expansion (Eq. 5), this condition becomes n + m = l, or equivalently, n – m must be a multiple of 3 to obtain a metallic nanotube, otherwise the nanotube is a semiconductor (Hamada et al., 1992). An armchair nanotube, for which n = m, is always a metal. This can be seen directly from Fig. 2(b): a line drawn along the arrow marked A always intersects the upper K point of the first Brillouin zone. The band gap of a semiconducting nanotube is easily calculated. Close to the K point, the function F(k) can be linearized with respect to the distance įk of the wave vector from the K point. To first order in įk, Eq. 4 simplifies in Er
3 2
H S r J 0 dCC G k
(6)
127
Figure 2. Planar development of the nanotube (n,m) for n = 5 and m = 3. In the rolledup structure, OC is the circumference of the nanotube, ș is the chiral angle. (b) Contour plot of the separation between the ʌ* and ʌ bands of graphene in reciprocal space. The ī point is at the center of the figure, the corners K of the first Brillouin zone are indicated by the black dots, the M points are indicated by the crosses. The thick lines across the drawing are discretization lines of the Bloch vector for the (5,3) nanotube, parallel to its axis. The two arrows indicate the axial direction of the armchair (A) and zigzag (Z) nanotubes.
where dCC is the CC bond length (0.14 nm). As can be seen from Fig. 2(b), the closest distance of the discretization lines to a K point for a semiconductor is one third the separation between these lines, which is the reciprocal of the nanotube radius. Setting įk = 2/3d with d the nanotube diameter in the above equation leads to the smallest possible energy of the ʌ* states and the largest possible value of the ʌ states. The separation between them is the band gap Eg
2J 0 dCC / d .
(7)
This formula is asymptotically correct for large d. The band structure of a nanotube can be obtained directly from Eq. 4 by restricting the wave vector k to the appropriate discretization lines and therefore making it a onedimensional good quantum number k (see Fig. 2). The calculations are straightforward in the case of armchair and zigzag nanotubes, and we find Elr,r (k ) rJ O 1 r 4cos(lS / n) cos(ka / 2) 4cos 2 ( ka / 2)
(8)
for the armchair nanotube (n,n), with a 3dCC the lattice parameter of graphene (translational axial period of the armchair tube) and
128 Elr,r (k ) rJ O 1 r 4cos(lS / n) cos( kc / 2) 4cos 2 (lS / n)
(9)
for the zigzag nanotube (n,0) where c = 3dCC is the axial period. In these expressions, l = 0, 1 ··· n – 1 labels the character of the wavefunction under the rotational symmetry group Cn that both of these nonchiral nanotubes possess. The ± sign in the subscript refers to the ± sign in front of ȖO (corresponding to the ʌ* and ʌ bands). The other ± comes from additional symmetry of the nanotube. For the armchair nanotube, all the bands have a twofold degeneracy (for instance, l = 1 is identical to l = n – 1), except for l = 0 and possibly l = n/2 when n is an even number. This is also true for the zigzag nanotubes. The band structure of an armchair (n,n) nanotube is shown in Fig. 3(a) for n = 10. The important characteristics is the presence of two branches that cross each other at the Fermi level (zero of energy) at 2/3 of the Brillouin zone extension. These two branches come from l = 0 in Eq. 8. Their wavefunctions are totally symmetric upon a rotation of 2ʌ/n about the nanotube axis. Their dispersion relation is E0, r (k ) rJ 0 2cos( ka / 2) 1 . The band structures of two zigzag (n,0) nanotubes, a semiconductor (n = 17) and a metal (n = 18), are shown in Fig. 3(b) and Fig. 3(c), respectively. The highest occupied band and the lowest unoccupied band of the semiconducting nanotube are doubly degenerate; they come from l = [n/3] and l = n – [n/3], where [x] means the nearest integer number to x. For the metallic (n,0) tube, two bands cross each other at the Fermi level at the ī point. These doublydegenerate bands correspond to l = n/3 and l = 2n/3 in Eq. 9. Their dispersion relation is En/ 3, r ( k ) r2J 0 sin(ka / 4) .
Figure 3. ʌelectron band structure of (a) the (10,10) nanotube, (b) the (17,0) nanotube, and (c) the (18,0) nanotube. The zero of energy corresponds to İʌ, which is where the Fermi level is located for undoped nanotubes.
129 In the metallic case, the branches that cross each other at the Fermi level EF have a nearly linear dispersion in the vicinity the Fermi wavevector. According to Eq. 6, their slope is dE / dk =vF r3J 0 dCC / 2 where, by definition, vF is the Fermi velocity. Each of these bands contributes a constant value 1/(S dE / dk ) to the density of states per unit length N(E). Multiplying by 4 (number of bands crossing EF) leads to N ( E ) 8 /(3SJ 0 dCC ) , independent on the nanotube radius. 2 Dividing N(E) by the number of atoms per unit length, 4S d /( 3dCC ) , yields the density of states per atom 2 3dCC S 2J 0 d
n( E )
(10)
which is valid for all metallic nanotubes for E § EF. The density of states is constant around the Fermi level and inversely proportional to the tube diameter d. Using Eq. 6 again, it is readily derived that the highest occupied and lowest unoccupied bands of a semiconducting nanotube (Fig. 3(b)) have a hyperbolic dispersion relation in the neighboring of the wave vector km where the band gap is located: Er (k ) H S r
E
/ 2 3J 0 dCC / 2 (k km ) 2 . 2
g
2
(11)
These relations are valid for all nanotubes, including the chiral ones.
Figure 4. ʌelectron density of states of (a) the (10,10) nanotube, (b) the (17,0) nanotube, and (c) the (13,6) nanotube. The calculations have been performed with Ȗ0 = 2.9 eV and İʌ = 0, which corresponds to the Fermi level for undoped nanotubes.
130 The density of states of three nanotubes is shown in Fig. 4. For the (10,10) nanotube (Fig. 4(a)), the plateau of density of states (Eq. 10) is clearly seen on both side of the Fermi level. The width of this plateau is three times larger than the corresponding band gap of a semiconductor nanotube having the same diameter, namely w 6J 0 dCC / d
(12)
The other two nanotubes in Fig. 4 are semiconductors with approximately the same diameter, with a band gap of about 0.6 eV. An essential characteristics of the density of states of nanotubes is the series of asymmetric spikes, called van Hove singularities, arising at all energies where a branch has a horizontal slope in the band structure. These singularities form intrinsic and easytoobserve fingerprints of the nanotube. Their positions can be measured by scanning tunneling microscopy (STS) (Wildoer et al., 1998; Odom et al., 1998; Kim et al., 1999), which sometimes makes it possible to assign a pair of wrapping indices to the nanotube, or simply to determine the nanotube diameter from Eq. 7 or Eq. 12 (Venema et al., 2000), depending on the metallic or semiconductor character of the tube. An optical absorption spectrum of singlewall carbon nanotubes is mainly determined by electronic transitions between van Hove singularities, where there is a large accumulation of states. These optical transition energies can be measured experimentally (Kataura et al., 1999). For light polarized parallel to the nanotube axis, transitions between symmetric states with respect to İʌ are allowed. Among these, the transition across the direct band gap of the semiconducting nanotubes has the lowest energy. It correspond to a first optical absorption band at E11S Eg  0.6 eV for usual nanotubes (those with diameter around 1.4 nm). The zonefolding approximation predicts another absorption band at about twice this energy E11S 2 Eg still for the semiconducting nanotubes. As for the metallic singlewall nanotubes, the transition between the van Hove singularities located on both sides of the metallic plateau in their density of states (see Fig. 4(a)) has an energy E11M w , which is typically 1.8 eV. At the time of writing, there is no real consensus on the actual value of the hopping interaction Ȗ0. Measurements of the dispersion relation of the two metallic crossing bands by STS in armchair nanotubes (Ouyang et al., 2002) gives Ȗ0 = 2.5 eV, whereas resonant Raman scattering, which probes the first optical transition energies, lead to Ȗ0 = 2.9 eV (Souza et al. 2004). The origin of the discrepancy between these two determinations is not known. Of course, the zonefolding approximation is only valid asymptotically for large nanotube diameter. But, even with more refined techniques such as nonorthogonal tight binding model, which incorporates the four valence electrons of carbon, the
131 optical transition energies measured experimentally are larger than the calculated ones by 0.3 eV (Popov, 2004). For not too small diameters, the correction to the bandstructure model is attributed to selfenergy and excitonic effects, which seem to partly compensate each other (Kane and Mele, 2004).
3. Transport properties The transport properties of a singlewall nanotube, either a metallic or a doped semiconductor, may deviate significantly from Ohm's law (Kong et al., 2001). Mesoscopic effects appear when the length of the nanotube between the contact electrodes (see Fig. 5) is smaller than the socalled coherence length. The coherence length is the average distance between two inelastic collisions with phonons, for instance. This distance increases by decreasing temperature simply because the number of thermally excited phonons is a decreasing function of temperature. For a metallic singlewall nanotube, the coherence length can be of the order of 1 µm at room temperature. Below that length, the electron wave function conserves memory of its phase, even if the electrons are elastically scattered by static defects and impurities. Even though, the resistance of the nanotube is not zero. The reason is that the nanotube offers only a few propagating states in comparison with the macroscopic contact electrodes that have many such states.
Figure 5. Thin wire connection (C) between two contact electrodes (L and R).
The conductance of the nanotube is proportional to the probability (transmission coefficient) that it transmits charges from one electrode to the other through its available electronic states ȥn. In a long, perfect nanotube, these are Bloch states corresponding to the energy bands shown in Fig. 3. The total current flowing from the left electrode to the right electrode is
132 I
³ ¦T (E) f n
L
( E EF eV )[1 f R ( E EF )]I n ( E )dE
n
³ ¦T (E) f n
R
( E EF )[1 f L ( E EF eV )]I n ( E )dE
(13)
n
where fR(E) and fL(E) are the FermiDirac distribution functions of the left and right electrodes, respectively, Tn(E) is the transmission coefficient for the state ȥn of the nanotube at the energy E, In+(E) and In(E) are the forward and backward currents per energy unit that state ȥn can carry. For onedimensional systems, it turns out that I n r ( E ) B2e / h , irrespective of the state (band) index n and energy E, provided E be located in the energy interval spanned by this state (Datta, 1995) (otherwise both Tn and In± vanish). Then, I
2e ¦ Tn ( E )[ f R ( E EF ) f L ( E EF eV )]dE h ³ n
(14)
At zero temperature, this formula further simplifies into I
2e E F ¦ Tn ( E )dE h ³EF eV n
(15)
At small bias, I 2e 2 / h V ¦ n Tn ( EF ) , which can be identified with the law I = GV, where the differential conductance G is given by G
2e 2 h
¦T (E n
F
)
(16)
n
Here the sum is over all nanotube electronic states n that encompass the Fermi energy EF and that can carry a current in one direction (also called conducting channels). Equation 16 is the wellknown LandauerBüttiker formula (Landauer, 1970). In the case of a perfect, infinite nanotube, Tn = 1 for all states, and G G0 M
(17)
where G0 = 2e2/h is the quantum of conductance and M is the number of branches that cross EF with a positive dE/dk slope in the band structure. The electrical resistance is therefore G 1 G01 / M 12.9 / M kȍ. It is lengthindependent (ballistic transport) and quantized due to the finite number M of conducting channels. For a metallic nanotube M = 2 when EF is close to the chargeneutrality energy İʌ. There are indeed two branches with positive slope in the band structure ("right'' or "left'' going branches) crossing İʌ, one at a positive k (shown in Fig. 3(a)), the other at the symmetric wave vector k. And
133 indeed, two quanta of conductance have been measured experimentally in some metallic nanotubes (Kong et al., 2001). When moving away from the Fermi energy, more bands are able to transport the current, an effect that gives a corresponding increase in G. In reality, since a nanotube is never perfect, the propagating electrons will be scattered by lattice defects, phonons, structural deformations of the nanotube or at the contacts. This leads to an unavoidable reduction of the transmission probability Tn(E) and in turn of the conductance. In practice, the resistance of a metallic nanotube can be significantly larger than 6.45 kȍ when there are poor coupling contacts with the external electrodes. The collision of electrons with phonons and defects is expected to increase the electrical resistance. Indeed, the resistance of a nanotube increases slightly with increasing temperature above room temperature, due to the backscattering of electrons by thermally excited phonons (Kane et al., 1998). At low temperature, the resistance of an isolated metallic nanotube also increases upon cooling (Yao et al., 1999). This unconventional behavior for a metallic system may be the signature of electron correlation effects first observed in SWNT ropes (Luttinger liquid, typical of onedimensional systems) (Bockrath et al., 1999). A single impurity, like B or N, in a metallic nanotube affects only weakly the conductivity close to İʌ (Choi et al., 2000). The same holds true with a StoneWales defect, which transforms four adjacent hexagons in two pentagons and two heptagons. Interestingly, even longrange disorder induced by defects like substitutional impurities has a vanishing backscattering cross section in metallic nanotubes, at least when EF is close to İʌ (Ando and Nakanishi, 1998; White and Todorov 1998). This is not true with doped semiconducting nanotubes, where scattering by longrange disorder is effective. A reduction of conductance will also happen when molecules (either chemisorbed or physisorbed molecules, i.e. dopants) interact with the tube (Meunier and Sumpter, 2005). For a nonperfect nanotube, it is clear that the conductance cannot simply be evaluated from counting the bands for a given electron energy but requires an explicit calculation of the transmission function. The difficulty arises from the fact that this type of calculation must be performed in an open system (nor finite or periodic), consisting of a conductor connected to the macroscopic world via two (or more) leads. Practically, the transmission coefficients can be evaluated efficiently using a Green's function and transfermatrix approach for computing transport in extended systems Buongiorno Nardelli, 1999), which can be generalized for multiterminal transport (Meunier et al., 2002). This method is applicable to any Hamiltonian that can be described with a localizedorbital basis.
134 A twoterminal system like in Fig. 5 can be partitioned into a left lead (L), the conductor (C), and a right lead (R). Assuming an orthogonal, localizedorbital basis, the Green function equation writes §H HL ¨ ¨ H CL ¨ 0 ©
H LC H HC H RC
0 · ¸ H CR ¸ H H R ¸¹
1
§ GL ¨ ¨ GCL ¨ 0 ©
GLC GC GRC
0 · ¸ GCR ¸ GR ¸¹
(18)
where H and G with appropriate indices refer to corresponding blocks of the Hamiltonian and Green'function, and İ is a complex energy (see below). From the Green functions, the selfenergies of the leads can be calculated in the form ȈL = HCLgLHLC and similarly for the right electrode, where gL = (İ  HL)1. The coupling Hamiltonian HLC is very shortranged thanks to the localizedorbital basis. As a consequence, the calculation of ȈL requires a few elements of the Green function gL of the perfect, periodic and semiinfinite lead. Using the selfenergies, the Green function of the conductor writes GC
H H C 6 L 6 R
The total transmission coefficient T calculated as a trace T
¦
1
(19)
T between the two leads can be
n n
Tr * L GCr * R GCa
1
(20)
and the current between the leads follows from Eq. 14. The coupling operator īL between the conductor and the left lead is expressed as a function of the advanced (a) and retarded (r) selfenergies, * L i 6 rL 6 aL , and similarly for the right electrode. In all these equation, İ = E ± iȘ where the arbitrarily small imaginary part Ș is added or subtracted to the energy E for the retarded or advanced Green's functions, respectively. Recent reports in the literature underline the dramatic role played by the interface between electrodes and the nanotube (Leonard and Tersoff, 2000). For instance, it has recently been shown that the electrical properties of junctions between a semiconducting carbon nanotube and a metallic lead dominate the overall electrical characteristics of nanotube based fieldeffect transistors (Appenzeller et al., 2002; Heinze et al., 2002; Martel et al., 2001; Meunier et al., 2002). In order to focus on the properties of the conductor itself, i.e., its intrinsic properties, one usually relies on modeling. Fortunately, theoretical methods conveniently offer the possibility to separate the effects of the intrinsic properties of the conductor from the effects of its connection to metallic leads. The effect of the heterogeneous contact between carbon nanostructures and metallic leads by seamlessly connecting the conductor to the electron reservoirs
135 by semiinfinite nanotube leads is illustrated in Fig. 6 for the case of a (5,5) nanotube. In this case, the transmission curve (solid line) is a steplike function that is merely a signature of the available bands in the system, each contributing for one channel of conductance (Eq. 17). When the conductor is connected to reservoirs by metallic leads, the solution of the Schrödinger equation of the perfect tube is no longer valid for the full system made up of the metallic leads and the nanotube portion; it follows that there is a reduction of the transmission probability (dashed line Fig. 6). In this case the interface between the device and the electrodes acts a scattering center.
Figure 6. Top: Electronic band structure of a perfect (5,5) carbon nanotube, where ka is the product of the Bloch wavevector k and the axial period a of the tube. The ʌ and ʌ* bands are crossing at 2/3 of the FBZ. The energy, in units of Ȗ0, is relative to the charge neutrality level. Bottom: Electron transmission in a (5,5) carbon nanotube, seamlessly connected to an identical nanotube in a bulklike fashion (full curve) and connected to heterogeneous metallic leads (dotted curve). The bulklike conductance curve corresponds to a mere counting of the allowed bands for any given energy, since the Bloch states of the leads are also solutions of the Schrödinger equation of the full system. When connected to heterogeneous metallic leads the Bloch states are no longer solution of the whole system, and the conductance can take any values, including noninteger ones.
136 4. Nanotube junctions The possibility of realizing seamless connections between singlewalled nanotubes with different helicities has stimulated a great interest for about ten years, because metalsemiconductor junctions with simple interfaces can be obtained in this way. Rectifying current properties were predicted for these systems (Service, 1996), like in a diode, which was further verified experimentally (Yao et al., 1999). A simple end to end connection between two different nanotubes, called an intramolecular junction, can be realized by the incorporation of an equal number of pentagons and heptagons. In the simplest case, this junction has just one pair of them (Dunlap, 1992; Saito et al., 1996). More complex structures can be formed, still by incorporating pentagons and heptagons in the otherwise perfect honeycomb arrangement of the carbon atoms. For instance, ideal T or Y junctions can be realized with an excess of six heptagons over the pentagons (Ferrer et al., 1999). Singlewall intramolecular junctions have been occasionally identified by AFM and STM in samples produced by laser ablation (Yao et al., 1999; Ouyang et al., 2001; Kim et al., 2003). Kinked nanotube junctions, which probably have their pentagon and heptagon at opposite sides, have been observed in a relatively large amount in catalytic CVD samples (Han et al., 1998) and also in arcdischarge samples (Hassanien et al., 2003). It is believed that intramolecular junctions can also be produced, perhaps in a controlled way, by electron irradiation of individual singlewall nanotubes (Hashimoto et al., 2004). If a twoterminal device like a metalsemiconductor intramolecular junction offers little interest for logical circuits compared to alltransistor devices, it presents the advantage of being a simple object to deal with. On the theoretical side, intramolecular junctions are ideal systems to consider for the study of the electronic structure of metal/semiconductor (Lambin et al., 1995; Chico et al., 1996; Charlier et al., 1996; Ferreira et al., 2000), metalmetal (Tamura and Tsukada, 1997), and semiconductor/semiconductor (Kim et al., 2003) interfaces, to investigate the Schottky barrier formation in such junctions (Odintsov, 2000; Yamada, 2002), and to calculate their transport properties (Saito et al., 1996; Chico et al., 1996b; Buongiorno Nardelli, 1999; Rochefort and Avouris, 2002). As an example, we discuss here the electronic properties of a straight connection between a metallic (5,5) nanotube and a semiconductor (10,0) nanotube (Rochefort and Avouris, 2002). This connection requires five pentagonheptagon pairs, distributed all around the structure, in such a way that the hybrid system conserves a fivefold symmetry that both constituents have. The atomic structure is shown in Fig. 7(a). The local density of states computed
137 with ʌ electrons around the rings of atoms labeled a, b … s are displayed in Figs. 7(b) and (c). As Fig. 7(b) demonstrates, the band gap rapidly forms as one proceeds into the semiconducting part of the junction, except that there is a welllocalized metalinduced gap state (MIGS) 0.28 eV below the İʌ level taken as zero of energy. This state exists between locations labeled b and persists in the metallic section up to ring n in Fig. 7(c). In the metallic part, backscattering of the wavefunctions whose energy falls within the band gap of the semiconductor lead to interferences typical of armchair nanotubes (Lambin et al., 1995; Meunier et al., 1998). These interferences lead to oscillations of the density of states in the region of the plateau of density of states, between 1.8 and +1.8 eV. Interestingly, the Landauer conductance calculated for this junction drops to zero in an interval of energy larger than the band gap of the (10,0) semiconducting part (Rochefort and Avouris, 2002). This is due to the fact that the wave functions of connected nanotubes have characters upon the C5v symmetry group of the junction that are not compatible in an energy window corresponding to the metallic plateau of the DOS of the (5,5) nanotube. As a consequence, the transmission probability vanishes there, like in the (6,6)(12,0) metalmetal system with sixfold symmetry (Chico et al., 1996b). The MIGS being below the Fermi level is occupied and is responsible for a charging effect, which amounts to 1.5 electron in the semiconducting part (rings ag) and 0.8 electron in the metallic part.
Figure 7. (a) Atomic structure of a straight and symmetric connection between (10,0) and (5,5) nanotubes. (b) and (c) Local ʌelectron densities of states averaged over the atoms forming the rings labeled a, b … s in (a).
138 By combining two intramolecular junctions in series, a quantum dot can be obtained when the metallic part is sandwiched between two semiconducting tubes (Chico et al., 1998). This system presents discrete peaks of conductance due to the finite number of electronic states localized in the metallic island (Lu et al., 2004). More interesting perhaps is the metal/semiconductor/metal structure, which may behave like a field effect transistor (Hyldgaard and Lundqvist, 2000). The conceptual simplicity of this device makes it attractive, at least for theoretical calculations. However, controlling the synthesis and growth of a double intramolecular junction of that sort would be extremely difficult. Another way to achieve the same effect, which could possibly be achieved experimentally, is by inducing a metalsemiconductor transformation by a local radial deformation of the central part of a metallic armchair nanotube (Lu et al., 2004). From the theoretical side, however, the structure of the squashed tube is more difficult to address than that of a perfect semiconducting tube sandwiched between two metallic ones.
Figure 8. (a) Atomic structure of a metalsemiconductormetal double junctions constructed with (12,0) and (11,0) zigzag nanotubes. (b) Local ʌelectron densities of states averaged over the atoms forming the rings labeled a, b … m in (a). Reproduced from another work (Triozon et al., 2005) of the authors published by IOP Publishing Ltd.
As an example of the latter system, we present here the case of a semiconducting (11,0) nanotube sandwiched between two semiinfinite metallic (12,0) tubes of helicity (12,0). Each (12,0)(11,0) junction includes a single pentagonheptagon pair aligned parallel to the axis. In the model shown in Fig. 8, the (11,0) section has 5 unitcells and length L = 2.1 nm. The local DOS, averaged on successive carbon rings of the (11,0)(12,0) junction, are plotted in
139 Fig. 8. Close to the interface, there is a finite DOS in the gap of the semiconductor due to the presence of MIGS's. The weight of these states decreases exponentially with the distance from the interface, with a decay length of order of 1 nm. Apart from the MIG states, there is another interesting feature in the local DOS of the semiconducting section: Several peaks are observed close to –1 and +1 eV, which can be attributed to discrete states confined between the two interfaces.
Figure 9. Variation of the conductance of the doublejunction (12,0)(11,0)(12,0) for three values of the length L of the semiconducting portion. Reproduced from another work of the authors (Triozon et al., 2005) published by IOP Publishing Ltd.
The energydependent conductance G(E) was calculated for different lengths L of the semiconducting section using Eq. 20, where the selfenergies and the Green functions were obtained by the socalled dissemination technique, as described elsewhere (Triozon et al., 2005). Variations of G with energy make sense if one considers shifting the Fermi level by a gate electrode and/or for transport under a large bias voltage. In both cases, conducting channels far from the charge neutrality point E = İʌ = 0 will contribute to the current. The results are shown in Fig. 9. For a short semiconducting section (L = 2.1 nm), electrons with energies falling in the gap can cross the semiconducting section. This is due to the overlap of the MIGS's from the two interfaces. This overlap is much smaller for L = 4.2 nm. Besides, out of the gap, we observe small oscillations of the conductance at E § 1 eV, corresponding to the peaks observed in the local DOS. These peaks are the signature of electronic states mostly localized in the semiconducting section, like in a quantum box. The density of states in the connected metallic (12,0) tubes is indeed small at
140 these energies, giving rise to quasilocalized states that provide conducting channel through the semiconducting island. When L increases, the separation energy between these discrete confined states gets smaller, roughly like 1/L, leading to faster oscillations of the conductance, especially near the edges of the gap. It is interesting also to remark the asymmetry of the conductance curve, which is due to the pentagons and heptagons that destroy the electronhole symmetry. ACKNOWLEDGEMENTS
The authors acknowledge M. Buongiorno Nardelli, V.N. Popov, and S. Roche for fruitful discussions. This research is supported in part by the IUAP project P5/01 "Quantum size effects in nanostructured materials" of the Belgian Science Policy Programming (Ph.L) and in part by the Division of Materials Science, U.S. Department of Energy under Contract No. DEAC0500OR22725 with UTBattelle, LLC at Oak Ridge National Laboratory (VM).
References Mintmire, J. W., Dunlap, B. I., and White, C. T., 1992, Are fullerene tubules metallic ?, Phys. Rev. Lett. 68:631634. Hamada, N., Sawada, S. I., and Oshiyama, A., 1992, New onedimensional conductors: graphitic microtubules, Phys. Rev. Lett. 68:15791581. Saito, R., Fujita, M., Dresselhaus, G., and Dresselhaus, M. S., 1992, Electronic structure of chiral graphene tubules, Appl. Phys. Lett. 60:22042206. Avouris, Ph., 2002, Carbon nanotube electronics, Chem. Phys. 281:429445. Connétable, D., Rignanese, G. M., Charlier, J. C., and Blasé, X., 2005, Room temperature Peierls distortion in small diameter nanotubes, Phys. Rev. Lett. 94:015503/14. Bockrath, M., Cobden, D. H., Lu, J., Rinzler, A. G., Smalley, E. E., Balents, L., and McEuen, P. L., 1999, Luttingerliquid behavior in carbon nanotubes, Nature 397:598601. Frank, S., Poncharal, Ph., Wang, Z. L., and de Heer, W. A., 1998, Carbon nanotube quantum resistors, Science 280:17446. Wallace, P. R., 1947, The band theory of graphite, Phys. Rev. 71:622634. Wildoer, J. W. G., Venema, L. C., Rinzler, A. G., Smalley, R. E., and Dekker, C., 1998, Electronic structure of atomically resolved carbon nanotubes, Nature 391:5962. Odom, T. W., Huang, J. L., Kim, Ph., and Lieber, Ch. M., 1998, Atomic structure and electronic properties of singlewalled carbon nanotubes, Nature 391:6264. Kim, Ph., Odom, T. W., Huang, J. L., and Lieber, C. M., 1999, Electronic density of states of atomically resolved singlewalled carbon nanotubes: Van Hove singularity and end states, Phys. Rev. Lett. 82:12251228. Venema, L. C., Meunier, V., Lambin, Ph., and Dekker, C., 2000, Atomic structure of carbon nanotubes from scanning tunneling microscopy, Phys. Rev. B 61:29912996. Kataura, H., Kumazawa, Y., Maniwa, Y., Umezu, I., Suzuki, S., Ohtsuka, Y., and Achiba, Y., 1999, Optical properties of singlewall carbon nanotubes, Synt. Metals 103:25552558.
141 Ouyang, M., Huang, J. L., and Lieber, C. M., 2002, Onedimensional energy dispersion of singlewalled carbon nanotubes by resonant electron scattering, Phys. Rev. Lett. 88:066804/14. Souza Filho, A. G., Chou, S. G., Samsonidze, G. G., Dresselhaus, G., Dresselhaus, M. S., An, L., Liu, J., Swan, A. K., Unlu, M. S., Goldberg, B. B., Jorio, A., Gruneis, A., and Saito, R., 2004, Stokes and antiStokes Raman spectra of smalldiameter isolated carbon nanotubes, Phys. Rev. B 69:115428/18. Popov, V. N., 2004, Curvature effects on the structural, electronic and optical properties of isolated singlewalled carbon nanotubes within a symmetryadapted nonorthogonal tightbinding model, New J. Phys. 6:17/117. Kane, C. L., and Mele, E. J., 2004, Electron interactions and scaling relations for optical excitations in carbon nanotubes, Phys. Rev. Lett. 93:197402/14. Kong, J., Yenilmez, E., Tombler, T. W., Kim, W., Dai, H., Laughlin, R. B., Liu, L., Jayanthi, C. S., and Wu, S. Y., 2001, Quantum interference and ballistic transmission in nanotube electron waveguides, Phys. Rev. Lett. 87:106801/14. Datta, S., 1995, Electronic Transport in Mesoscopic Systems, Cambridge University Press, Cambridge, 1995, pp. 1377. Landauer, R., 1970, Electrical resistance of disordered onedimensional lattices, Phil. Mag. 21:863867. Kane, C. L., Mele, E. J., Lee, R. S., Fischer, J. E., Petit, P., Dai, H., Thess, A., Smalley, R. E., Verschueren, A. R. M., Tans, S. J., andDekker, C., 1998, Temperaturedependent resistivity of singlewall carbon nanotubes, Europhys. Lett. 41:683688. Yao, Z., Postma, H. W. Ch., Balents, L., and Dekker, C., 1999, Carbon nanotube intramolecular junctions, Nature 402:273276. Choi, H. J., Ihm, J., Louie, S. G., and Cohen, M. L., 2000, Defects, quasibound states, and quantum conductance in metallic carbon nanotubes, Phys. Rev. Lett. 84:29172920. Ando, T., and Nakanishi, T., 1998, Impurity scattering in carbon nanotubes  Absence of back scattering, J. Phys. Soc. Japan 67:17041713. White, C. T., and Todorov, T. N., 1998, Carbon nanotubes as long ballistic conductors, Nature 393:240242. Meunier, V., and Sumpter, B. G., 2005, Amphoteric doping of carbon nanotubes by encapsulation of organic molecules: electronic properties and quantum conductance, J. Chem. Phys., in press. Buongiorno Nardelli, M., 1999, Electronic transport in extended systems: Application to carbon nanotubes, Phys. Rev. B 60:78287833. Meunier, V., Buongiorno Nardelli, M., Bernholc, J., Zacharia, T., and Charlier, J. C., 2002, Intrinsic electron transport properties of carbon nanotube Yjunctions, Appl. Phys. Lett. 81:52345236. Leonard, F., and Tersoff, J., 2000 Role of Fermilevel pinning in nanotube Schottky diodes, Phys. Rev. Lett. 84:46934696. Appenzeller, J., Knoch, J., Derycke, V., Martel, R., Wind, S., and Avouris, Ph., 2002, Fieldmodulated carrier transport in carbon nanotube transistors, Phys. Rev. Lett. 89:126801/14. Heinze, S., Tersoff, J., Martel, R., Derycke, V., Appenzeller, J., and Avouris, Ph., 2002, Carbon nanotubes as Schottky barrier transistors, Phys. Rev. Lett. 89:106801/14. Martel, R. V., Lavoie, C., Appenzeller, J., Chan, K. K., Tersoff, J., and Avouris, Ph., 2001, Ambipolar electrical transport in semiconducting singlewall carbon nanotubes, Phys. Rev. Lett. 87:256805/14. Service, R. F., 1996, Mixing nanotube structures to make a tiny switch, Science 271:1232. Dunlap, B. I., 1992, Connecting carbon tubules, Phys. Rev. B 46:19331936.
142 Saito, R., Dresselhaus, G., and Dresselhaus, M. S., 1996, Tunneling conductance of connected carbon nanotubes, Phys. Rev. B 53:20442050. Melchor Ferrer, S., Khokhriakov, N. V., and Savinskii, S. S., 1999, Geometry of multitube carbon clusters and electronic transmission in nanotube contacts, Molec. Eng. 8:315344. Ouyang, M., Huang, J. L., and Lieber, Ch. M., 2001, Atomically resolved singlewalled carbon nanotube intramolecular junctions, Science 291:97100. Kim, H., Lee, J., Kahng, S. J., Son, Y. W., Lee, S. B., Lee, C. K., Ihm, J., and Kuk, Y., 2003, Direct observation of localized defect states in semiconductor nanotube junctions, Phys. Rev. Lett. 90:216107/14. Han, J., Anantram, M. P., Jaffe, R. L., Kong, J., and Dai, H., 1998, Observation and modeling of singlewall carbon nanotube bend junctions, Phys. Rev. B 57:1498314989. Hassanien, A., Holzinger, M., Hirsch, A., Tokumoto, M., and Venturini, P., 2003, Ropes of carbon nanotubes intramolecular junctions, Synthetic Metals 137:12031204. Hashimoto, A., Suenaga, K., Gloter, A., Urita, K., and Iijima, S., 2004, Direct evidence for atomic defects in graphene layers, Nature 430:870873. Lambin, Ph., Fonseca, A., Vigneron, J. P., Nagy, J. B., and Lucas, A. A., 1995, Structural and electronic properties of bent carbon nanotubes, Chem. Phys. Lett. 245:8589. Chico, L., Crespi, V. H., Benedict, L. X., Louie, S. G., and Cohen, M. L., 1996, Pure carbon nanoscale devices: Nanotube heterojunctions, Phys. Rev. Lett. 76:971974. Charlier, J. C., Ebbesen, T. W., and Lambin, Ph., 1996, Structural and electronic properties of pentagonheptagon pair defects in carbon nanotubes, Phys. Rev. B 53:1110811113. Ferreira, M. S., Dargam, T. G., Muniz, R. B., and Latge, A., 2000, Local electronic properties of carbon nanotube heterojunctions, Phys. Rev. B 62:1604016045. Tamura, R., and Tsukada, M., 1997, Conductance of nanotube junctions and its scaling law, Phys. Rev. B 55:49914998. Odintsov, A. O., 2000, Schottky barrier in carbon nanotube heterojunction, Phys. Rev. Lett. 85:150153. Yamada, T., 2002, Modeling of kinkshaped carbonnanotube Schottky diode with gate bias modulation, Appl. Phys. Lett. 80:40274029. Chico, L., Benedict, L. X., Louie, S. G., and Cohen, M. L., 1996b, Quantum conductance of carbon nanotubes with defects, Phys. Rev. B 54:26002606. Rochefort, A., and Avouris, Ph., 2002, Quantum size effects in carbon nanotube intramolecular junctions, Nano Lett. 2:253256. Meunier, V., Henrard, L., and Lambin, Ph., 1998, Energetics of bent carbon nanotubes, Phys. Rev. B 57:25912596. Chico, L., Lopez Sancho, M. P., and Munoz, M. C., 1998, Carbonnanotubebased quantum dots, Phys. Rev. Lett. 81:12781281. Liu, H., and Tao, Y., 2003, Size effect of quantum conductance in singlewalled carbon nanotube quantum dots, Eur. Phys. J. B 36:411418. Hyldgaard, P., and Lundqvist, B. I., 2000, Robust nanosized transistor effect in fullerenetube heterostructure, Solid State Commun. 116:569573. Lu, J. Q., Wu, J., Duan, W., and Gu, B. L., 2004, Metaltosemiconductor transition in squashed armchair carbon nanotubes, Appl. Phys. Lett. 84:42034205. Triozon, F., Lambin, Ph., and Roche, S., 2005, Electronic transport properties of carbon nanotube based metal/semiconductor/metal intramolecular junctions, Nanotechnology 16:230233.
ELECTRONIC TRANSPORT IN CARBON NANOTUBES AT THE MESOSCOPIC SCALE SYLVAIN LATIL Department of Physics, Facultés Universitaires NotreDame de la Paix, 61 Rue de Bruxelles, B 5000 Namur, Belgium FRANÇOIS TRIOZON DRT/LETI/DIHS/LMNO, CEA, 17 Rue des Martyrs, F 38054 Grenoble Cedex 9, France STEPHAN ROCHE DSM/DRFMC/SPSMS/GT, CEA, 17 Rue des Martyrs, F 38054 Grenoble Cedex 9, France
Abstract. This chapter deals with the electronic transport in carbon nanotubes under the influence of a static disorder. The approach used is the Kubo formalism that is based on fluctuationdissipation theory developed in real space. Keywords: electronic transport; Kubo formalism; defects
1. Introduction and motivation A carbon nanotube (CNT) is a onedimensional graphitic structure that can be either metallic or semiconductors, owing to its geometry (Saito et al., 1998). In addition to this unusual electronic structure, CNTs also exhibit remarkable quantum transport properties,1 which present them as serious candidates for emerging nanoelectronics. First basic components have been synthesized:
______ 1
Electronic transport in carbon nanotubes seems to be very robust under static disorder (Franck et al., 1998). Theoretical works have nicely explained such unusual property (White and Todorov, 1998). *To whom correspondence should be addressed. Sylvain Latil; email: [email protected]
143 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 143–165. © 2006 Springer. Printed in the Netherlands.
144 junction nanodiode (Zhou et al., 2000) and logic gate (Derycke et al., 2001). Physical studies of such nanojunctions are full of promise.2 In order to tailor the electronic properties of CNTs, the Fermi level can be tuned by chemical doping, which can be achieved by different approaches: intercalation of alkali metal atoms (Fischer, 2002), substitution of boron or nitrogen in carbon nanotubes. Carbon nanotubes doped either with nitrogen or boron (Terrones et al., 1996; Czerw et al., 2001; Carroll et al., 1998) substitutions have been synthesized. When substituting a carbon atom, the boron atom (resp. nitrogen) acts as an acceptor (resp. donor) impurity in the nanotube, as revealed by thermopower measurements (Choi et al., 2003). On the other hand, the chemical sensing capabilities of carbon nanotubesbased devices appear very promising (Star et al., 2003a). Indeed, beyond the standard covalent bonding (Bahr et al., 2001) that is responsible for a significant perturbation of the CNT, functional groups can specifically be attached to the carbon nanotube surface by physisorption (Star et al., 2001b). On the other hand, the ʌstacking of ʌconjugated organic molecules on the nanotube sidewalls is an example of noncovalent functionalization involving a weaker binding energy (Tournus, et al., 2005). The induced scattering is thus expected to be low and moleculardependent in opposition to the electrochemical covalent functionalization. Consequently, a good knowledge of the CNTs reactivity and of the impact of the physisorption on the quantum transport is needed to guarantee their potential application as chemical sensors. One notes that the chemical reactivity of nanotube/metal interface is also an important issue of concern, when investigating the behaviors and performances of nanotubesbased field effects transistors. By means of a suited chemical treatment of contacts, optimization of charge injection properties was endeavored (Auvray et al., 2004). In this lecture, we present an approach to study the electronic transport in disordered carbon nanotubes (and any 1D material). The method we use and the related Kubo formalism are discussed in Section 2. Section 3 presents the first application of this formalism to atomic substitutions of boron in nanotubes. We will extract the electronic transport quantities of nanotubes as a function of the density of dopants. In Section 4, we present results on nanotubes with a random coverage of ʌconjugated molecules and establish that the HOMOLUMO gap of the adsorbed molecular entities is a crucial parameter for backscattering.
______
2 Electronic properties of nanotubes (with the tightbinding or zone folding Hamiltonian), junctions and the Landauer formalism are presented in the Chapter by P. Lambin, V. Meunier and F. Triozon.
145 2. The Kubo formalism in real space The conductivity of a bulk material of volume ȍ is defined for frequency Ȧ as the tensorial ratio between the applied field and the resulting electronic current: j(Ȧ) = ı(Ȧ)E(Ȧ). Since we will perform calculations on 1D materials, oriented along the zaxis, only the diagonal element will be taken into account: jz(Ȧ) = ı(Ȧ)Ez(Ȧ). The Kubo approach is a technique to calculate linear responses of materials (optical, electric, etc.). It is based on the fluctuationdissipation theorem that establishes a correspondence between the dissipative outofequilibrium response (namely, the conductivity) and the fluctuations at the equilibrium (the correlation function of the charge carrier velocities). In this situation, the Kubo formula of conductivity reads (Kubo et al., 1985)
2S e2 = f f ( E ) f ( E =Z ) ˆ G ( E H)V ˆ ˆ G (E H ˆ =Z ) º dE Tr ¬ª V (1) z z ¼ : ³f =Z ˆ is the Hamiltonian operator, V ˆ is the operator for the electronic where H z velocity along the zaxis and f(E) is the Fermi distribution function. The DC conductivity corresponds to the limit Ȧ = 0. Using the property
V (Z )
lim
Z o0
f ( E ) f ( E =Z ) =Z
wf wE
G ( E EF )
(2)
and after a Fourier transform (Roche and Mayou, 1997), the diagonal conductivity writes in real space:
ª1 º e2 n( EF )lim « 'Z 2 (t ) » E t o0 t ¬ ¼
V DC
(3)
where n(EF) is the density of states per unit of volume and 'Z 2 (t ) is the E measure of the electronic quadratic spreading at energy E.3 It is defined like: 'Z 2 (t )
E
2 ˆ Zˆ (t ) Zˆ (0) º Tr ªG ( E  H) ¬« ¼» ˆ Tr ª¬G ( E  H) º¼
(4)
______ 3
In term of fluctuationdissipation theory, the movement of an electronic wave packet due to an external electric field (i.e., dissipation under non equilibrium conditions) is related to the spreading of the wave packet without any external electric field (i.e., fluctuations at the equilibrium).
146 where Zˆ (t ) is the position operator along the zaxis, written in the Heisenberg representation for the time t. In view of simplifications, we will modify equation (4). Firstly, using the timereversal symmetry and the properties of the Trace, it is straightforward to demonstrate that
2 ˆ Zˆ (t ) Zˆ (0) º Tr ªG ( E  H) »¼ ¬«
ª Zˆ , uˆ (t ) º ¬ ¼
A(t )
ˆ A(t ) º Tr ª¬ A (t )G ( E  H) ¼ ˆ ˆ (t ) uˆ (t ) Zˆ Zu
(5a) (5b)
where Zˆ is the position operator in the Schrödinger representation and ˆ t / = ) is the usual evolution operator. Secondly, the Traces in uˆ(t ) exp(iH equation (4) are approximated by expectation values on wavepackets (Triozon et al., 2002) that are treated as randomphase states:4 Tr >...@ o wp ... wp
and the spreading (4) can finally be rewritten as 'Z 2 (t )
ˆ A(t ) wp wp A (t )G ( E  H) . ˆ wp wp G ( E  H)
E
(6)
Equation (6) is now suitable for order O(N) numerical techniques and calculation of the transport properties are possible. We leave the technical details in Appendices A and B and turn now to the physical discussion about this approach. In fact, the quadratic spreading (6) is a key quantity as it is directly related to the diffusion coefficient (or diffusivity) DE (t )
Z 2 (t )
E
1 t
(7)
whose time dependence fully determines the transport mechanism. It is worth also to define the electronic spreading Z E (t )
Z 2 (t )
E
tDE (t ).
(8)
______ 4
Random phase states are expanded on all the orbitals n of the basis set and defined like 1 N wp ¦ exp(2iSD (n)) n , N n1 where D (n) is a random number in the [0,1] range. An average over 10 to 20 random phases states is sufficient to calculate the expectation values.
147 Three main categories of transport mechanisms can be interpreted, as illustrated in Fig. 1. The ballistic regime represents the ideal case of a perfect 1D lattice. Electrons travel through the systems without any scattering: the correlation of their group velocities is perfect after any time, and both DE (t ) and Z E (t ) are linear functions, with a slope equal respectively to v 2f and v f (Fig. 1, left). Such electronic transport is characterized by a quantum conductance that does not decrease with length, and a steplike curve for G ( E ) . The diffusive regime is characterized by a saturation of DE (t o f) , owing to identify a relaxation time W ( E ) , after what DE (t ) is constant and Z E (t ) ~ t (Fig. 1, center). For typical times longer than Ĳ, a semiclassical interpretation allows the recovery of the Einstein formula for conductivity (Ashcroft and Mermin, 1976) like
V DC # e 2 n( EF ) DE . F
The most useful tools to characterize a diffusive motion at energy E are the electrons group velocity v E and the elastic meanfreepath le ( E ) le ( E )
v E W (E)
(9a)
v
DE (t ) t
(9b)
and E
with t W ( E ) .
The conductance of diffusive 1D systems (whose length is longer than le ) linearly decreases with their length. The localized regime is a pure quantum regime that does not allow any semiempirical explanation. In very disordered systems, the electronic wave function cannot propagate because of destructive interferences. Conducting electrons are then "trapped" onto limited regions and the diffusivity behaves like ~1/t. Spreading Z E (t ) reaches an asymptotic value (we can identify to the size of the "trapping'' regions) that is called the localization length ] ( E ) (Fig. 1, right)
] ( E ) lim Z E (t ) . t of
(10)
148
Figure 1. Typical behavior of the diffusion coefficient DE (t ) and the spreading Z E (t ) , given by equations (7) and (8) for the three characteristic regimes: ballistic (a), diffusive (b) and localized (c).
149 Let's turn now to an important remark. All the dynamic of the system will ˆ operator that will be set up following the tight binding be driven by the H 2 approach. Since the Hamiltonian is constructed with the presence of static disorder (e.g., randomly located defects), the elastic scattering will be taken automatically onto account. Our aim is to quantify this scattering by calculating its length scales: elastic meanfreepath, localization length, etc. But, as we can see, the system remains Hamiltonian: its total energy is conserved. In this ˆ possesses an infinite lifetime. In reality, power situation, an eigenstate of H dissipation must occur somewhere: in the contacts or in the system (due to an energy exchange with the phonons), and conducting electrons loose their phase due to any inelastic scattering that does not conserve the energy. The resulting phase coherence length LM cannot be computed within this approach and has to be seen as an arbitrary parameter. We will always consider that LM is much longer than the size of the studied system, i.e., that the decoherence process is negligible in the system and the dissipation occurs only in the contacts. In this quantum coherent regime, from the Kubo formula, and since the system is onedimensional, the generic conductance of a device of length Ldev writes: G ( E , Ldev ) 2e 2 n( E )
DE (W dev ) , Ldev
(11)
where n(E) is the electronic DOS per unit length and W dev ( E ) is the time spent by the wave packet (at the considered energy E) to spread over a distance equal to Ldev . Following the fluctuationdissipation framework, it is also equivalent to the time needed for an electron to travel through the device.5 3. Boron substitutions In this section, the generic transport properties of borondoped nanotubes such as conduction mechanisms, meanfreepaths and conductance scalings, are computed as a function of the density of dopants. Electronic calculations have been carried on (10,10) nanotubes, with 2500 cells and periodic boundary conditions (105 atoms). To use equation (6), the zone folding (ZF) Hamiltonian has been used. This method is an orthogonal tightbinding (TB) approach, which takes into account only one orbital pA per atom, and describes correctly the usual electronic properties of pristine carbon
______ 5
This approach is equivalent to the Landauer formula, assuming reflectionless contacts (Fisher and Lee, 1982; Datta, 1995).
150 nanotubes.2 The presence of the gap, due to the local curvature, in the band structure of "metallic" nanotubes (Ouyang et al., 2001) is not predicted by ZF calculations, but since this work addresses doped systems, a truly metallic behavior should be recovered. Within this framework, the Hamiltonian operator writes as follows ˆ H
N
ª
¦ «H n 1
¬
nn
º n n ¦ H np n p » p ¼
(12)
where the first sum runs on all the pA orbitals in the system while the second runs on the first neighbors of the n site. In expression (12), the matrix elements Hnn are the onsite energies and Hnp are the hopping integrals.
Figure 2. Electronic properties of a boron substitution in a graphene sheet. a) Isodensity plot of the last (half) occupied band, at ī point calculated with DFTLDA. Plane is shifted from the graphitic sheet (0.625 Å). Electronic density is mainly distributed on the B atom, up to the third neighbor. b) The renormalized atoms used our model are labeled: 1st, 2nd and 3rd neighbors of the boron (B) atom. c) Comparison between the electronic band structures of the doped graphene sheet calculated with DFTLDA (left) and our modified TB model (right) with the renormalized onsite energies (see text).
151
Figure 3. a) Electronic diffusion in a (0.1 %) Bdoped (10,10) CNT. The DOS illustrates in (a) the characteristic acceptor peek located below the Fermi level (here EF = 0). b) The diffusivity DE(t) given by equation (7), as a function of time, (b) for the three ranges of energy, indicated by arrows in (a). The diffusion coefficient for energy E1 is ten times magnified. c) Diffusivity DE(t) as a function of energy for a time t = 200ƫ/Ȗ0 for the same Bdoped CNT (solid line) and a pristine CNT (dashed line).
In order to apply this ZF technique to borondoped carbon nanotubes, the electronic properties in the vicinity of an atomic substitution need to be accurately described. Tightbinding methods (like the ZF) are usually not suitable for predicting charge transfer, especially, in the chosen case of heteroatomic substitutions. However, by adding a corrective electrostatic potential to the onsite energies, this technique can handle electric fields, charge transfer or electric dipole moments. The correction (added to the onsite energies) is usually calculated by a selfconsistent loop on the electric charge (Krzeminski et al., 2001). Unfortunately, as the present O(N) technique does not allow to use such a selfconsistent scheme, an alternative approach is explained below for the case of a Bdoped CNT. In our modified ZF model, the values of the matrix elements in equation (12) are supposed to vary depending on the involved species. These elements are labeled İC, İB,… (for the onsite energies Hnn) and ȖCC, ȖBC,…(for the hopping integrals Hnp). Practically, these parameters were defined by fitting the
152 ZF band structure on DFTLDA calculations.6 In this abinitio approach, standard normconserving pseudopotentials were used, and the cutoff energy for the plane waves expansion was set to Ecut = 30 Ha. Since the curvature of the graphene sheet is neglected in the ZF approximation, the LDA calculations were performed on flat "graphenelike" systems. At first, the electronic structure of a supercell containing 31 carbon atoms and a single B atom was studied within a spinaveraged LDA approach, in order to simulate the electronic states of a doped carbon system in the vicinity of the B impurity. As shown in Fig. 2a, the electronic density ȡ(r) for the last (half) occupied band is distributed only on the pA orbitals for atoms located close to the impurity, up to the third neighbor of the B atom. This localization of the HOMOLUMO band allows considering that the correction on the onsite energies affects carbon atoms only up to the third neighbors of the impurity. Moreover, this result suggests that the hopping integral between sites will not be affected by the charge transfer in assumption that the ʌ atomic orbitals are not polarized by the local electric field. In addition, the boron atom is supposed to be "carbonlike"7, i.e., ȖCC = ȖBC = Ȗ. The geometry of the model is presented in Fig. 2b, where the boron and the renormalized carbon atoms are labeled. In this situation, only 6 parameters need to be adjusted: the unique hopping integral Ȗ, the carbon and boron onsite energies İC and İB, and the renormalized carbon onsite energies İ3, İ2, and İ1 (resp. third, second and first neighbors of the boron atom, as shown in Fig. 2b). These adjustments were performed using least square energy minimization scheme between LDA and ZF band structures. At first, the LDA electronic structure of an isolated graphene sheet was used to fit the hopping. Its value was kept further as a constant. As only a low density of boron atoms in a graphene sheet is considered and given that this supercell is supposed to be in electronic equilibrium with the surrounding nanotube, the chemical potentials (Fermi energies) of the two subsystems have to be equal. Since the Fermi energy of a graphene sheet (or a nanotube described with ZF technique) is İC, this leads to İC = EF,supercell = EF,CNT. The band structure obtained with the optimal parameters is compared to the LDA band structure in Fig. 2c. The optimal hopping integral is Ȗ =2.72 eV and the Fermi level EF = İC. The onsite energies are İB = +2.77
______
6 The DFTLDA calculations were done with the ABINIT code. This program is a common project of the Université Catholique de Louvain, Corning Incorporated, and other contributors (http://www.abinit.org). Electronic states are expanded on a plane wave basis set (energy cutoff is 35 Hartree) and standard normconserving pseudopotentials have been used (Troullier and Martins, 1991). 7 Since the electronic density for a C and a B atom at a distance d = 1.42 Å is roughly the same, this approximation is valid.
153 eV, İ1 = –0.16 eV, İ2 = +0.21 eV, and İ3 = –1.56 eV. At last, the whole spectrum is shifted to have EF = 0 eV. As shown in Fig. 3a, the density of states (DOS) of a (0.1%) Bdoped (10,10) CNT exhibits the typical acceptor peak (E1) in agreement with previous abinitio studies (Choi et al., 2000). On Fig. 3b, three different transport regimes are illustrated for a (0.1%) Bdoped nanotube. At Fermi energy (EF), the diffusion coefficient saturates at long times where D(EF,t) ĺ D0 ~ levF, indicating a diffusive regime. In contrast, at the resonance energy (E1) of the quasibounded states, the diffusivity exhibits a ~ 1/t behavior, typical of a strong localization regime. Finally at energy E2 above the Fermi level, the electronic conduction remains nearly insensitive to dopants, as shown by the quasiballistic diffusion law. In Fig. 3c, the energydependent diffusivity clearly manifests such asymmetry in conduction.
Figure 4. Scaling of the mean free path le at the Fermi level for a Bdoped (n,n) nanotube. Left: in the case of a (10,10) nanotube with various boron concentrations, le behaves like the inverse of the doping rate. Right: for a fixed concentration of B atoms, le is roughly a linear function of the diameter.
From a physical point of view, the relevant information is that a low density of dopants yields diffusive regimes, with a meanfreepath decreasing linearly with dopant concentration following Fermi golden rule (Fig. 4, left), and increasing linearly with nanotube diameter (Fig. 4, right) following theoretical predictions based on Andersonlike8 disorder modeling (White and Todorov, 1998). Moreover, in very good agreement with experimental data (Liu et al.,
______
8 In the Anderson model, the disorder is not properly created by localized scattering centers but acts as a random modulation of the onsite energies. It then can be seen as a "white noise", affecting equally the whole energy spectrum (Anderson, 1958).
154 2002), from our calculations, we estimate meanfreepaths in the order of 175– 275 nm for borondoped nanotubes with diameters in the range 17–27 nm, and 1% of doping.
Figure 5. Quantum conductance of a device made from a single (10,10) nanotube containing 0.1% of boron impurities. The conductance is plotted as a function of the energy, for different lengths of the device. A decrease of the conductance is observed near the energies of donor states, and near the Van Hove singularities.
155 The main remarkable conductance features for Bdoped (10,10) nanotubes are illustrated in Fig. 5. As the typical length Ldev increases from the nanometric scale (~10 nm) to the mesoscopic scale (~ 1 µm), the quantum interference effects beyond the diffusive regime are correspondingly enhanced. A strong asymmetric damping of the electronic conductance follows. Moreover, our results show that the generic conductance G(E,Ldev) of BCNT at Fermi level exhibit a positive derivative with respect to energy. Similar study on nitrogen doped nanotubes indicates that localized states around the nitrogen substitution, are responsible for backscattering above the Fermi level (Latil et al., 2004). Such a difference between N and Bdoped CNTs suggests that the thermopower measurements9 originate from a diffusion mechanism only (Choi et al., 2003). Gatedependent studies on individual undoped carbon nanotubes have recently revealed a quantum connection between conductance modulations and thermopower (Small et al., 2003). Similar studies on chemically doped SWNTs would be desirable to confirm these theoretical predictions. 4. Adsorption of ʌconjugated molecules In this section, the effect of molecular physisorption on the diffusion mechanisms of metallic CNTs is theoretically investigated. As in Section 3, the results will be achieved by combining DFT calculations with tightbinding models. The focus is made on the evaluation of the scattering strength and the resulting transport length scales produced by a random coverage of ʌconjugated hydrocarbon molecules (benzene (C6H6) and azulene (C10H8)). In order to adjust the parameters of the semiempirical Hamiltonian, and since the interaction between the CNT and each molecule is geometrydependent, an ab initio study of the adsorption has been carried out prior to the TB analysis (Tournus et al., 2005). Assuming that the adsorption on a CNT or on graphene are equivalent, the atomic model is simplified when using a 5 x 5 graphene supercell with one adsorbed benzene or azulene molecule, instead of the cylindrical structure.
______ 9
The thermopower S gives important information about the sign of charge carriers and the conductance modulations close to the Fermi level:
S
S 2 k B2T § dG · ¨ ¸ 3eG ( EF ) © dE ¹ E
(Mott formula) EF
where G(E) is the conductance at Fermi level, kBT the temperature energy scale.
156
Figure 6. The electronic structures of a graphene 5 x 5 supercell with one adsorbed ʌconjugated molecule. Top: the HOMO eigenstate of the isolated benzene C6H6 molecule (shown with a dashed line in the left hand frame) mixes up with the band structure of the graphene layer, to form a complex band structure (central frame). Bottom: an azulene C10H8 molecule is physisorbed on the graphene sheet; the mixing also occurs for HOMO and LUMO eigenstates. Although a weaker dipersion is observed than for the benzene case, the smaller gap of azulene will allow a perturbation of the CNT electronic diffusion at the Fermi level (here EF = 0). Note that in both cases, our TB model gives an excellent description (right frame) compared to the abinitio calculation.
157
Figure 7. Schematic view of the graphenemolecule interaction described by equation (14). The distance d and the angle ĳ are drawn. b) The modified Hückel model for the azulene molecule. The onsite energies İ are corrected in order to take screening into account and the hopping integrals are Ȗ = –2.63 eV and Ȗ' = –2.08 eV, respectively.
Here, electronic calculations have been performed with the density functional theory within its local density approximation to predict the optimal geometry, and extract its electronic structure.10 From the optimized geometries11 (displayed in the center of Fig. 6), the LDA band structures are plotted for both isolated and interacting systems. The interaction acts as a mixing of the molecular single states with the underlying band structure, resulting in hybrid eigenstates with low group velocity. However, due to the relatively large gaps of the isolated molecules (5.217 eV and 2.089 eV for benzene and azulene, respectively) only weak disturbance is induced around the Fermi energy, and the linear band dispersion is preserved. Following the same procedure as in Section 3, the zone folding model is used for defining the Hamiltonian of the nanotube. The ʌconjugated molecules are treated by the Hückel model (Salem, 1966) and the tightbinding Hamiltonian of the complete system, containing the interaction between CNT and Nmol molecules, reads Nmol
ˆ H ˆ ˆ ªˆ º H CNT ¦ ¬ H mol ( M ) V( M ) ¼
(13)
M 1
ˆ where H CNT is the usual zone folding Hamiltonian, equivalent to equation (12) ˆ and H mol ( M ) is the Hückel hamiltonian related to the Mth molecule. The
______ 10
The AIMPRO code (Briddon and Jones, 2000) has been used with standard normconserving pseudopotentials (Bachelet et al., 1982) and a 3 x 3 x 1 kpoints sampling for the Brillouin zone integrations. The basis used is sufficient to obtain correct adsoption curves (C atoms contain five s and plike and three dlike orbitals, and H atoms contain four s and plike orbitals). 11 Equilibrium distances (energies of adsorption) are found to be 3.25 Å (25.1 kJ.mol1) for the benzene, and 3.28 Å (37.6 kJ. mol1) for the azulene molecules.
158 ˆ V(M) term in equation (13) corresponds to the interaction between the Mth molecule and the CNT, and is expressed as follow: ˆ M) V(
§ d G · ¸n m l ¹
¦ ¦ E cos(M ) exp ¨©
mCNT mM
(14)
Such coupling term was optimized to accurately reproduce the interaction between shells in a multiwalled CNT (Lambin et al., 2000). In Eq. (14), d corresponds to the distance between sites, while ĳ is the angle defined in Fig. 4a. The parameters related to the interaction are: į = 3.34 Å, l = 0.45 Å and ȕ = –0.36 eV.
Figure 8. a) DOS of the (10,10) CNT with a benzene density coverage of 16.3%. The HOMO molecular state is located by an arrow. b) Same as in a) but for an azulene density coverage of 11.5%. HOMO and LUMO levels are identified by arrows. c) Timedependent diffusion coefficient (at Fermi level) for the C6H6@CNT system, showing quasiballistic behavior (solid line). Ballistic conduction for pristine CNT case is also reported (dashed line). d) Timedependent diffusion coefficient for the C10H8@CNT structure, showing saturation at large times (diffusive regime).
159 The tightbinding parameters (onsites energies İ and hopping integrals Ȗ) are adjusted to reproduce the band structure of graphene sheet and the energy levels of isolated benzene and azulene molecules, calculated within LDA. A standard procedure have been used to set the parameters for the graphene or the CNT: İCNT = 0 eV, ȖCNT = –2.56 eV, as well as for the benzene molecule: İbenz = +0.411 eV, Ȗbenz = –2.61 eV. Due to an inhomogeneous charge distribution along the azulene molecule, the TB model has to be refined to properly account for screening effects. By adding an electrostatic correction, proportional to the net charge on each carbon atom (calculated with LDA), renormalized onsite energies for the azulene molecule have been deduced. The onsites and the adjusted hopping integrals are given in Fig. 4b. Within this approach, the TB value of the azulene HOMOLUMO gap is 2.198 eV. Finally, the band structure computed with this modified TB model is compared to the previous ab initio results. As presented in Fig. 6, this reparametrized semiempirical model gives an excellent description of the electronic states for both benzene and azulene adsorption cases. The density of states (DOS) of a (10,10) carbon nanotube, with random coverage of adsorbed molecules is calculated within this framework. Fig. 8 presents the DOS for a nanotube with benzene coverage density of 16.3% (such density corresponds to the ratio of adsorbed mass of the molecule over CNT mass) and with azulene coverage density of 11.5%.Since the coupling intensity is weak, the DOS plotted in Fig. 8 displays the reminiscent discrete molecular levels, slightly enlarged by the mixing with the underlying continuum of ʌ or ʌ*bands. Peaks arising from molecular HOMO and LUMO levels are labeled. In the case of the benzene adsorption, the DOS is weakly affected at charge neutrality point, and the computed diffusion coefficient shows almost no deviation from the linear scaling in time (ballistic regime). Elastic backscattering induced by benzene molecules is thus vanishingly low at Fermi level (also shown in Fig. 8), and does not produce sufficient deviation to determine an elastic mean free path below ~100 µm. In contrast, in the case of azulene molecules, the HOMO level is located in the close vicinity of the last occupied Van Hove singularity of the CNT. Although the total DOS remains basically unaffected around the Fermi energy, the azulene adsorption turns out to impact more significantly on the intrinsic electronic current. Indeed, as shown in Fig. 8, the diffusion coefficients of propagating electrons at Fermi level depart more strongly from the ballistic regime. The nonlinear time dependence of DE(t), calculated with equation (7) allows to extract the elastic relaxation time Ĳ and, knowing the Fermi velocity vF, the corresponding electronic mean free path le. Fig. 9 presents the evolution
160 of le as a function of the azulene coverage density, both at Fermi level (charge neutrality point) and for an energy closer to the HOMO resonance [E = –0.2Ȗ § –0.54 eV].
Figure 9. Electronic mean free path of a C10H8@CNT nanostructure versus the density of physisorbed molecules and for two different Fermi level positions. The origin of the error bars is the use of the random phase states for the expectation values in equation (6).
AKNOWLEDGEMENTS
We acknowledge fruitful discussions and collaborations with Prof. Didier Mayou, Prof. JeanChristophe Charlier, Prof. Philippe Lambin, Dr. Xavier Blase and Prof. Malcolm I. Heggie. S. L. acknowledges financial support from the Belgian Francqui Fundation for Scientific Research. A. THE RECURSION TECHNIQUE
As we have established in Section 2, the main issue of our numerical technique is to compute expectation values, like the local density of states on a given ket M related to the retarded Green's function:
161
1 1 ª º M » limH o0 « Im M ˆ S E iH H ¼ ¬
ˆ M M G ( E H)
(15)
We use the realspace recursion technique, which avoids diagonalization or inversion of the Hamiltonian and has a numerical cost which grows linearly with the system size (order N) (Haydock et al., 1972). Starting from a normalized ket M , an orthonormal basis { M0 , M1 ,…} is constructed from the recurrence relations:
ˆ M H 0
ˆ M H n
a0 M0 b0 M1
(16a)
an Mn bn1 Mn1 bn Mn1
(16b)
ˆ M , hence it is a real with M0 = M . At any step n, an is given by Mn H n number. Then bn and Mn 1 are uniquely determined by the normalization condition for Mn 1 and the choice of a real positive bn. In this basis, the Hamiltonian is tridiagonal. The Green's function element mentioned above is simply the first diagonal element of the inverse of the tridiagonal matrix: § E iH a0 ¨ ¨ b0 ¨ ¨¨ ©
b0 E iH a1 b1
b1 E iH a2 .
· ¸ ¸ .¸ ¸ . ¸¹
It is obtained by a continued fraction expansion:
M0
1 M ˆ 0 E iH H
1 z a0
(17)
b02 z a1
b12 ...
With Nrecurs recursion steps, an energy resolution of order of W/Nrecurs is obtained for the density of states, where W is the energy bandwidth of the system. In the singleband case, an and bn converge and the continued fraction is terminated by the approximation aN+1 = a and bN+1 = b. A more sophisticated termination can be used in a multiband system, as detailed in (Roche, 1999).
162 B. COMPUTATION OF A(t ) wp BY CHEBYCHEV POLYNOMIALS EXPANSION
ˆ uˆ (t ) º on This section presents the method used to apply the operator A(t ) ª¬ Z, ¼ a state vector M , in view of solving numerically the equation (6), with time steps ǻt. The concept consists in expanding the evolution operator uˆ ('t ) on polynomials of the Hamiltonian operator like: uˆ ('t )
f
ˆ ¦ c ('t )Q (H) n
(18)
n
n 0
Practically, sums are limited to Npoly (hence, the exponential in the evolution operator is approximated by a polynomial of order Npoly and Qn are the Chebychev polynomials, which follow the simple recursive rules
Hˆ a Q (H)ˆ bQ
ˆ bQn1 (H)
n 1
n
ˆ ˆ 1 ,. Q1 (H) Q0 (H)
1 b 2
ˆ (H)
Hˆ a
(19a)
(19b)
The real numbers a and b are parameters defining a density function on the mathematical domain [a – 2b, a +2b]
1
N (E)
§ Ea· 2S b 1 ¨ ¸ © 2b ¹
2
(20)
ˆ within the It is then crucial to choose a and b to fit all the spectrum of H domain. The expansion coefficients cn(ǻt) in equation (18) are then the scalar product like
cn (t )
i
³ N ( E)Q (E)exp( = E't )dE n
(21)
and can be computed. To simplify the equations (18) and (19), we will introduce the two sets of kets
Dn En that obeys the recursion rules
ˆ M Qn (H)
(22a)
ˆ ˆ º ª Z, ¬ Qn (H) ¼ M
(22b)
163
b D n1
Hˆ a D
b D n1
(23a)
ˆ a M H 2
(23b)
ˆ ˆ aº E b E ª Z,H n 1 ¬ ¼ n
(24a)
1
M , D1
D0
n
b
and
b En1
1
ˆ ˆºM ª Z,H ¼ b 2¬
0 u M , E1
E0
(24b)
ˆ 12 and H ˆ. They can be calculated from the starting ket M by applying Z Once we know D n and En , the application of the operators on M is simply
uˆ ('t ) M
Npoly
¦ c ('t ) D n
(25a)
n
n 0
ˆ ˆ ª Z, º ¬ u ('t ) ¼ M
Npoly
¦ c ('t ) E n
n
(25b)
n 0
Finally, when the equations (25) are solved for a given time t,13 the following step t + ǻt is calculated by applying the Chebychev expansion on solution at time t like
uˆ (t 't ) M
ªˆ ˆ º ¬ Z, u (t 't ) ¼ M
uˆ ('t )uˆ (t ) M
ªˆ ˆ ºˆ ªˆ ˆ º ˆ ¬ Z, u ('t ) ¼ u (t ) M u ('t ) ¬ Z, u (t )¼ M
(26a)
(26b)
and so on. The evolution of A(t ) M and uˆ (t ) M can be calculated step by step from any starting M .
______
12 The choice of an orthogonal basis set (within the TB approach) simplifies the action of the ˆ p G Z . position operator, assuming that n Z np n 13 ˆ uˆ (t )º M have already been calculated. This means that the two vectors uˆ(t ) M and ª¬ Z, ¼
164 References Anderson, P. W., 1958, Absence of Diffusion in Certain Random Lattices, Phys. Rev. 109: 1492. Ashcroft, N. W., and Mermin, N. D., 1976, Solid State Physics, Saunders College, Philadelphia. Auvray, S., Borghetti, J., Goffman, M. F., Filoramo, A., Derycke, V., Bourgoin, J.P., and Jost, O., 2004, Carbon nanotube transistor optimization by chemical control of the nanotubemetal interface, Appl. Rev. Lett. 84: 5106. Bachelet, G. B., Haman, D. R., and Schlüter, M., 1982, Pseudopotentials that work: From H to Pu, Phys. Rev. B 26: 4199. Bahr, J. L., Yang, J. P., Kosynkin, D. V., Bronikowski, M. J., Smalley, R. E., and Tour, J. M., 2001, Functionalization of Carbon Nanotubes by Electrochemical Reduction of Aryl Diazonium Salts: A Bucky Paper Electrode, J. Am. Chem. Soc. 123: 6536. Briddon, P. R. and Jones, R., 2000, LDA calculations using a basis of gaussian orbitals, Phys. Status Solidi B 217: 131. Carroll, D. L., Redlich, P., Blase, X., Charlier, J.C., Curran, S., Ajayan, P., Roth, S., and Rühle, M., 1998, Effects of Nanodomain Formation on the Electronic Structure of Doped Carbon Nanotubes, Phys. Rev. Lett. 81: 2332. Choi, H. J., Ihm, J., Louie, S. G., and Cohen, M. L., 2000, Defects, Quasibound States, and Quantum Conductance in Metallic Carbon Nanotubes, Phys. Rev. Lett. 87: 2917. Choi, Y.M., Lee, D.S., Czerw, R., Chiu, P.W., Grobert, N., Terrones, M., ReyesReyes, M., Terrones, H., Charlier, J.C., Ajayan, P. M., Roth, S., Carroll, D. L., and Park, Y.M., 2003, Nonlinear Behavior in the Thermopower of Doped Carbon Nanotubes Due to Strong, Localized States, Nano Lett. 3: 839. Czerw, R., Terrones, M., Charlier, J.C., Blase, X., Foley, B., Kamalakaran, R., Grobert, N., Terrones, H., Tekleab, D., Ajayan, P. M., Blau, W., Ruhle, M., and Carroll, D. L., 2001, Identification of Electron Donor States in NDoped Carbon Nanotubes, Nano Lett. 1: 457. Datta, S., 1995, Electronic Transport in Mesoscopic Systems, Cambridge Univesity Press, Cambridge. Derycke, V., Martel, R., Appenzeller, J., and Avouris, P., 2001, Carbon Nanotube Inter and Intramolecular Logic Gates, Nano Lett. 1: 453. Fischer, J. E., 2002, Chemical doping of singlewall carbon nanotubes, Acc. Chem. Res. 35: 1079. Fisher, D., and Lee, P. A., 1982, Relation between conductivity and transmission matrix, Phys. Rev. B 23: 6852. Frank, S., Poncharal, P., Wang, Z., and de Heer, W., 1998, Carbon nanotube quantum resistors, Science 280: 1744. Haydock, R., Heine, V., and Kelly, M., 1972, Electronic structure based on the local atomic environment for tightbinding bands, J. Phys. C: Solid State Phys. 5: 2845. Krzeminski, C., Delerue, C., Allan, G., Vuillaume, D., and Metzger, R. M., 2001, Theory of electrical rectification in a molecular monolayer, Phys. Rev. B 64: 085405. Kubo, R., Toda, M., and Hashitsume, N., 1985, Statistical Physics II, Springer, Berlin. Lambin, P., Meunier, V., and Rubio, A., 2000, Electronic structure of polychiral carbon nanotubes, Phys. Rev. B 62: 5129. Latil, S., Roche, S., Mayou, D., and Charlier, J.C., 2004, Mesoscopic Transport in Chemically Doped Carbon Nanotubes, Phys. Rev. Lett. 92: 256805. Liu, K., Avouris, P., Martel, R., and Hsu, W. K., 2002, Electrical transport in doped multiwalled carbon nanotubes, Phys. Rev. B 62: 161404. Ouyang, M., Huang, J.L., Cheung, C. L., and Lieber, C. M., 2001, Energy Gaps in Metallic SingleWall Carbon Nanotubes, Science 292: 702.
165 Roche, S., 1999, Quantum transport by means of O(N) realspace methods, Phys. Rev. B 59: 2284. Roche, S., and Mayou, D., 1997, Conductivity of Quasiperiodic Systems: A Numerical Study, Phys. Rev. Lett. 79: 2518. Saito, R., Dresselhaus, G., and Dresselhaus, M. S., 1998, Physical Properties of Carbon Nanotubes. Imperial College Press, London. Salem, L., 1966, The MolecularOrbital theory of ʌconjugated systems. Benjamin, Reading. Small, J. P., Perez, K. M., and Kim, P., 2003, Modulation of Thermoelectric Power of Individual Carbon Nanotubes, Phys. Rev. Lett. 91: 256801. Star, A., Han, T. R., Gabriel, J. C., Bradley, K., and Gruner, G., 2003a, Interaction of Aromatic Compounds with Carbon Nanotubes: Correlation to the Hammett Parameter of the Substituent and Measured Carbon Nanotube FET Response, Nano Lett. 3: 1421. Star, A., Liu, Y., Grant, K., Ridvan, L., Stoddart, J. F., Steuerman, D. W., Diehl, M. R., Boukai, A., and Heath, J. R., 2003b, Noncovalent SideWall Functionalization of SingleWalled Carbon Nanotubes, Macromolecules 36: 553. Terrones, M., Benito, A. M., MantecaDiego, C., Hsu, W. K., Osman, O. I., Hare, J. P., Reid, D. G., Terrones, H., Cheetham, A. K., Prassides, K., Kroto, H. W., and Walton, D. R. M., 1996, Pyrolytically grown BxCyNz nanomaterials: nanofibres and nanomaterials, Chem. Phys. Lett. 257: 576. Tournus, F., Latil, S., Heggie, M. I., and Charlier, J.C., 2005, ʌstacking interaction between carbon nanotubes and organic molecules, Phys. Rev. B 72: 075431. Triozon, F., Vidal, J., Mosseri, R., and Roche, S., 2002, Quantum dynamics in two and threedimensional quasiperiodic tilings, Phys. Rev. B 65: 220202. Troullier, N., and Martins, J. L., 1991, Efficient pseudopotentials for planewave calculations, Phys. Rev. B 43: 1993. White, C. T., and Todorov, T. N., 1998, Carbon nanotubes as long ballistic conductors, Nature 393: 240. Zhou, C., Kong, J., Yenilmez, E., and Dai, H., 2000, Modulated Chemical Doping of Individual Carbon Nanotubes, Science 290: 1552.
WAVE PACKET DYNAMICAL INVESTIGATION OF STM IMAGING MECHANISM USING AN ATOMIC PSEUDOPOTENTIAL MODEL OF A CARBON NANOTUBE
GÉZA I. MÁRK,* LEVENTE TAPASZTÓ, LÁSZLÓ P. BIRÓ Research Institute for Technical Physics and Materials Science, H1525 Budapest, P.O.B. 49, Hungary ALEXANDRE MAYER Département de Physique, Facultés Universitaires NotreDame de la Paix, 61 Rue de Bruxelles, B5000 Namur, Belgium
Abstract. Scanning tunneling microscopy (STM) is capable of providing information on both the geometrical and electronic structure of a carbon nanostructure. Utilizing a pseudopotential model of a nanotube we investigated the threedimensional wave packet (3D WP) tunneling process through an STM tip – nanotube – support model for different atomic arrangements representing metallic and semiconducting nanotubes.
Keywords: Scanning tunneling microscopy; carbon nanotube; local pseudopotential
Scanning Tunneling Microscopy (STM) images of nanosystems contain both the effect of the geometry and the electronic structure. To investigate geometrical effects in the STM imaging mechanism, formerly we performed wave packet dynamical simulations1,2 for jellium models of STM tip – carbon nanotube (CNT) – support tunnel junctions. We were able to explain some geometrical effects frequently seen in STM images. In the present paper we extend the 3D WP calculation to include the details of the atomic structure. A local one electron pseudopotential3 was used for the CNT. The STM tip of 0.5 nm radius and the conductive support surface was modeled by constant potential jellia.2 The time development of a Gaussian WP approaching the tunnel junction from inside the tip bulk was calculated by numerically solving the time dependent 3D Schrödinger equation. The timedependent probability *To whom correspondence should be addressed. Géza I. Márk; email: [email protected]
167 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 167–168. © 2006 Springer. Printed in the Netherlands.
168 density is visualized by snapshots of a constant density surface. In the panel t = 0.0 fs of Fig. 1 the initial WP is shown – still in the tip bulk region. At t = 0.6 fs the WP has already penetrated into the tip apex region. At t = 1.8 fs the WP has already tunneled from the tip into the tube and is beginning to flow around the tube circumference. The WP flows along the CC bonds with negligible density at the centers of the hexagons. At t = 3.02 fs the WP already has flown around the tube circumference and is beginning to spread along the tube. Spreading is faster for the metallic tube than for the semiconducor tube. These preliminary results show that the 3D WP tunneling simulation is a useful tool for interpreting experimental data and predicting the likely behavior of nanodevices.
Figure 1. Details of tunneling of a wave packet approaching the STM junction from the tip bulk and tunneling into the nanotube. Constant probability density surface is shown. Upper row: (10,0) tube, lower row: (6,6) tube. The isosurface is clipped at the presentation box boundaries.
This work was partly supported by OTKA grants T 43685 and T 43704 in Hungary and the Belgian PAI P5/01 project on "Quantum size effects in nanostructured materials''. The calculations were done on the Sun E10000 supercomputer of the Hungarian IIF.
References 1 2
3
G. I. Márk, L. P. Biró, and J. Gyulai, Simulation of STM images of 3D surfaces and comparison with experimental data: carbon nanotubes, Phys. Rev. B 58, 1264512648 (1998). G. I. Márk, L. P. Biró, and Ph. Lambin, Calculation of axial charge spreading in carbon nanotubes and nanotube Y junctions during STM measurement, Phys. Rev. B 70, 115423/111 (2004). A. Mayer, Band structure and transport properties of carbon nanotubes using a local pseudopotential and a transfermatrix technique, Carbon 42(10), 20572066 (2004).
CARBON NANOTUBE FILMS FOR OPTICAL ABSORPTION EVA KOVATS,* ARON PEKKER, SANDOR PEKKER, FERENC BORONDICS, KATALIN KAMARAS Research Institute for Solid State Physics and Optics, 2933 KonkolyThege M. Street H1121 Budapest, Hungary
Abstract. Ultrathin, transparent, and pure nanotube films have great importance in the optical studies. The samples are freestanding, which eliminates the difficulties caused by the various substrates in widerange optical measurements. We prepared uniform homogenous thin films from holedoped and functionalizated singlewall carbon nanotubes (SWNTs) with a simple process. We examined the spectra from the infrared to the ultraviolet range on the same sample. The vibrational transitions of the sidewall groups indicated the chemical composition. The changes in the electronic properties were followed simultaneously by Nearinfrared/Visible/Ultraviolet (NIR/VIS/UV) spectroscopies. Keywords: carbon nanotube thin film; optical spectroscopy; absorption
Carbon nanotubes have unique electronic properties, which have been widely investigated on isolated SWNTs as well as on bundled nanoropes. Here we describe widerange optical studies on SWNTs prepared as freestanding films. Freestanding nanotube films were produced by vacuum filtration of dilute nanotube suspensions.1 After the dissolution of the filter, the films were transferred to a frame with a hole. The films were thin enough to be optically transparent (100200 nm) and wide enough to have a sufficient surface area (0.51 cm2). The optical measurements were made in the spectral range from 600 to 30000 cm1. For the mid and nearinfrared regions we used Bruker a IFS 66v spectrometer. The visible and UV data were obtained on a Jasco V550. The molecular vibrations can be detected up to 4000 cm1 and above this wavenumber the electronic transitions can be observed. *To whom correspondence should be addressed. Eva Kovats, Research Institute for Solid State Physics and Optics, 2933 KonkolyThege M. Street H1121 Budapest, Hungary; email: [email protected]
169 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 169–170. © 2006 Springer. Printed in the Netherlands.
170 We examined two chemically modified nanotube samples. In the purification process, the SWNTs were treated with concentrated nitric acid, which causes charge transfer doping. In the absorption spectra, we detected increase of the M00 band intensity. We explain this with the increase of the number of free charge carriers. The other sample, butylated SWNTs, was prepared by reductive alkylation. 2,3 In the 28003100 cm1 region, several CH modes of butyl sidegroups were detected. The occupation of the van Hove singularities decreased due to less Selectrons.4 The slight change is due to the small degree of functionalization (1,8%, Ref. [3]). We measured the transmittance of pure nanotube films and calculated the optical conductivity.5 The simultaneous optical transparency and DC conductivity make nanotube films ideal materials for transparent electrodes.1 In summary, we prepared transparent freestanding nanotube films, which made possible widerange optical absorption measurements on the same sample. The change of the electronic properties was detectable. For example, in the case of ionic doping, the M00 absorption bands of electronic transitions increased because of more free electrons; the effect of functionalization causes the decrease of the S11, S22, M11 bands because of less Selectrons. Because of the high DC conductivity and high transmittance, these films are ideal candidates for transparent electrodes.
References 1. Z. Wu, Z. Chen, X. Du, J. M. Logan, J. Sippel, M. Nikolou, K. Kamaras, J. R. Reynolds, D. B. Tanner, A. F. Hebard, and A. G. Rinzler, Transparent, conductive carbon nanotube films, Science 305, 12731276 (2004). 2. S. Pekker, J.P. Salvetat, E. Jakab, J.M. Bonard, and L. Forro, Hydrogenation of Carbon Nanotubes and Graphite in Liquid Ammonia, J. Phys. Chem. B 105, 79387943 (2001). 3. F. Borondics, E. Jakab, M. Bokor, P. Matus, K. Tompa, S. Pekker, Reductive Functionalization of Carbon Nanotubes, Fullerenes, Nanotubes and Carbon Nanostuctures 13, 375 (2005) 4. K. Kamarás, M. E. Itkis, H. Hu, B. Zhao, R. C. Haddon, Covalent Bond Formation to a Carbon Nanotube Metal, Science 301, 1501 (2003). 5. F. Borondics, K. Kamaras, Z. Chen, A. G. Rinzler, M. Nikolou, D. B. Tanner, Wide range optical studies on transparent SWNT films, Electronic properties of synthetic nanostructures, AIP Conference Proceedings 723, 137140 (2004).
INTERSUBBAND EXCITON RELAXATION DYNAMICS IN SINGLEWALLED CARBON NANOTUBES C. GADERMAIER,* C. MANZONI, A. GAMBETTA, G. CERULLO, G. LANZANI ULTRAS – INFM, Dipartimento di Fisica, Politecnico di Milano, P. za L. da Vinci 32, 20133 Milan, Italy E. MENNA, M. MENEGHETTI Department of Chemical Sciences, University of Padova, 35131 Padova, Italy
Abstract. We study excited state dynamics of semiconducting singlewalled carbon nanotubes (SWNTs) by twocolor pumpprobe experiments. We find a time constant of 40 fs for the intersubband energy relaxation from EX2 to EX1.
Keywords: pumpprobe spectroscopy; carbon nanotubes; excitons
Timeresolved photoemission,1 pumpprobe,24 and fluorescence spectroscopy5 reveal that semiconducting SWNTs excited to their lowest excitonic state (EX1) show a fast recovery to the ground state due to nonradiative decay channels. Our study complements these findings by resolving the relaxation from the second subband exciton (EX2) to EX1. Transient oscillations due to coherent excitation of the radial breathing mode are seen in some of the time traces. The setup used consists of an amplified Ti:sapphire laser driving two noncollinear optical parametric amplifiers (NOPAs) which generate pulses tunable in the near IR and visible. We studied samples of SWNTs grown by the high pressure carbon monoxide procedure, functionalized with PEGamine and embedded in polymethylmethacrylate: films were cast on glass cover slips. Figure 1 shows 'T/T dynamics at two different pump and probe energies, 0.9 eV and 2.1 eV, respectively. The two pump energies are resonant with the EX1 and EX2 transitions. We propose the following simple assignment of the various signals to rationalize the observations. The positive signals are due to photobleaching of *To whom correspondence should be addressed. Christoph Gadermaier; email: [email protected]
171 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 171–172. © 2006 Springer. Printed in the Netherlands.
172
'T/T (arb. units)
the EX1 and EX2 transitions respectively. The PA at 2.1 eV is attributed to a transition between the first and third conduction/valence subbands as conjectured by Korovyanko et al. (Ref. [6]). 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 100
0
100
200
300
400
pumpprobe delay (fs)
Figure 1. Transient differential transmission of SWNT sample at 0.9 eV pump and 0.9 eV probe (open circles), 2.1 eV probe (closed circles), 2.1 eV pump and 0.9 eV probe (open squares) and 2.1 eV probe energies (closed squares). Solid lines show fits to the data.
Excitation of either EX1 or EX2 leads to an instantaneous bleaching of the respective transition. The relaxation of EX1 to the ground state occurs with a time constant of 400 fs, in good agreement with previous measurements4 EX2, on the other hand, relaxes towards EX1 with a time constant7 of 40 fs. This is responsible for the fast decay of the EX2 bleaching concomitant with a buildup of both the EX1 bleaching and the EX1EX3 absorption.
References 1. 2.
3.
4.
5. 6.
T. Hertel and G. Moos, Electronphonon interaction in singlewall carbon nanotubes: A timedomain study, Phys. Rev. Lett. 84, 50025005 (2000). Y.C. Chen, N. R. Raravikar, L. S. Schadler, P. M. Ajayan, Y.P. Zhao, T.M. Lu, G.C. Wang, and X.C. Zhang, Ultrafast optical switching properties of singlewall carbon nanotube polymer composites at 1.55 Pm, Appl. Phys. Lett. 81, 975977 (2002). H. Han, S. Vijayalakshmi, A. Lan, Z. Iqbal, H. Grebel, E. Lalanne, and A. M. Johnson, Linear and nonlinear optical properties of singlewalled carbon nanotubes within an ordered array of nanosized silica spheres, Appl. Phys. Lett. 82, 14581460 (2003). O. J. Korovyanko, C.X. Sheng, Z.V. Vardeny, A. B. Dalton, and R. H. Baughman, Ultrafast spectroscopy of excitons in singlewalled carbon nanotubes, Phys. Rev. Lett. 92, 017403 (2004). F.Wang, G. Dukovic, L.E. Brus, and T.F. Heinz, Timeresolved fluorescence of carbon nanotubes and its implication for radiative lifetimes, Phys. Rev. Lett. 92, 177401 (2004). C. Manzoni, A. Gambetta, E. Menna, M. Meneghetti, G. Lanzani, and G. Cerullo, Intersubband exciton relaxation dynamics in singlewalled carbon nanotubes, Phys. Rev. Lett. 94, 207401 (2004).
PECULIARITIES OF THE OPTICAL POLARIZABILITY OF SINGLEWALLED ZIGZAG CARBON NANOTUBE WITH CAPPED AND TAPERED ENDS O. V. OGLOBLYA Kyiv National Shevchenko University, Kyiv 01033, Ukraine G. M. KUZNETSOVA Scientific Center of Radiation Medicine, Kyiv 04050, Ukraine
Abstract. Computer modeling of the linear optical polarizability of singlewalled carbon nanotubes (SWCNTs) of the zigzag type (n,0) with different diameter capped at one end by half of fullerene and tapered at the other end so that it can be connected with different (n,n) armchair nanotubes is carried out. The calculations were carried out within the SuSchriefferHeeger model. It is shown that the localized states demonstrate the nonlinear aspects of excited states in that system. It is found that the molecules with different radius have strong oscillating dependence of the optical polarizability on the incident light energy. The effect of the length, capping of the tube end, and the Coulomb repulsion on the optical polarizability spectrum is studied.
Keywords: optical polarizability; carbon nanotube; SuSchriefferHeeger model
The SuSchriefferHeeger (SSH) model (Ma and Yuan, 1998) was used in the numerical calculations of the optical and electronic properties of an individual SWCNT and its modifications. We introduced new parameters [ D 2 /( Kt ) , u U / t , and v V / t . The values ȟ = 0.32 and t = 3.2 eV were used in the calculations as in a previous work (Prylutsky and Ogloblya, 2004). Within the independent electron approximation and sumoverstates approach, the linear polarizability is expressed as 2 2 D (Z ) ¦ nocc., punocc. Pnp P pn 2H pn (H pn Z 2 ) /[(H pn Z 2 )2 * 2Z 2 ] where Pnp is the dipole momentum. To account for the radiative widening of electron levels, we used ī = 32 meV. The main attention was devoted to the spatially averaged polarizability D ( D xx 2  D yy 2  D zz 2 ) / 3 . 173 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 173–174. © 2006 Springer. Printed in the Netherlands.
174 We investigated zigzag (15,0) nanotube, capped at the ends with caps made from C180 fullerene. The results of the computations of Į for nanotubes with length 6 and 8 periods are depicted in Figs. 1a and 1b, left. As we can see from these figures, the uncapped nanotube has additional lowenergy peaks that are due to dangling bonds at the opened ends of nanotubes. The length decrease or the uncapping of the nanotube shifts the peaks of the polarizability towards the high energy region and suppresses their values.
Figure 1. Left: polarizability of (a)open (15,0) and (b)capped at both ends nanotubes with length of 6 (solid line) and 8 (dotted line) periods (u = 0.3, v = 0.15). Right: polarizability of cappedtapered nanotube with u = 0; v = 0 (solid line), u = 0.3; v = 0.15 (dashed line), u = 0.6; v = 0.3 (dotted line).
We also considered the tapered configuration of the closing fragment (Saito et al., 1996). These authors describe a general way to connect nanotubes with different diameter by a knee configuration with a heptagonpentagon defect. We used this procedure and built a knee that connects a capped (10,0) nanotube with length 8 with a (2,2) nanotube, which was not completed and four bonds remained dangling. The results are depicted in Fig. 1, right, which demonstrates the influence of the Coulomb repulsion. A splitting of the main peak of the polarizability can be observed, which is larger for higher values of the Coulomb repulsion. References Ma, J., and Yuan, R. K., 1998, Electronic and optical properties of finite zigzag carbon nanotubes with and without Coulomb interaction, Phys. Rev. B 57:93439348. Prylutsky, Yu. I., and Ogloblya, O. V., 2004, Computer modelling of the optical absorption spectrum of singlewalled carbon nanotube bundles, Ukr. J. Phys. 49:A17A20. Saito, R., Dresselhaus, G., and Dresselhaus, M. S., 1996, Tunneling conductance of carbon nanotubes, Phys. Rev. B 53:20442050.
THIRDORDER NONLINEARITY AND PLASMON PROPERTIES IN CARBON NANOTUBES A. M. NEMILENTSAU,* A. A. KHRUTCHINSKII, G. YA. SLEPYAN, S. A. MAKSIMENKO Institute for Nuclear Problems, Belarus State University, Bobruiskaya 11, 220050 Minsk, Belarus
Abstract. Nonlinear optical properties in singlewall carbon nanotubes (CNTs) illuminated by an external electromagnetic field were investigated using quantum kinetic equations for the density matrix of the ʌelectrons. In the regime of weak driving field, these equations were solved by the perturbation method and the spectra of the thirdorder polarizability were calculated. In the case of highintensive driving field, numerical simulation of the processes in the timedomain was carried out. The ʌplasmons excitation was analyzed. Keywords: carbon nanotubes; thirdorder polarizability; plasmon
We investigated the nonlinear optical effects, which take place in an infinitely long rectilinear singlewall carbon nanotube (CNT) exposed to an external electromagnetic field. To calculate the nonlinear response of CNT, the quantum kinetic equations for the ʌ electrons motion in singleelectron approximation were obtained.1 We assumed that the CNT is illuminated by a normally incident electromagnetic field, polarized along the tube axis. To avoid the destruction of the CNT by the driving field, its strength should be smaller than the atomic field strength. The driving field was assumed to be uniform over the CNT. Then its wavelength should be much greater than the CNT crosssectional radius. In the regime of weak driving field, the quantum kinetic equations were solved by the perturbation method and the thirdorder polarizability ( 3) 3Z ; Z , Z , Z of the singlewall CNT was calculated. The spectra of this D zzzz polarizabillity were plotted in the frequency range from 1.35 to 1.9 eV for four different types of the CNTs: armchair (8,8), metallic zigzag (15,0), and semiconductor zigzag (16,0) and (38,0). The obtained spectra of the thirdorder *To whom correspondence should be addressed. A. M. Nemilentsau, Institute for Nuclear Problems, Belarus State University, Bobruiskaya 11 220050, Minsk, Belarus; email: [email protected]
175 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 175–176. © 2006 Springer. Printed in the Netherlands.
176 polarizability contain a number of resonant lines whose position and magnitude are strongly dependent on the CNT type. However, in spite of this fact, there are several spectral maxima whose location does not depend on the tube radius. They are the resonant lines caused by the plasmon with respect to the third harmonic of the driving field frequency. Our result is in a good agreement with the experiment on the thirdorder harmonic generation in the solution comprising CNTs illuminated by the fundamental and second harmonics of the ND:YAG laser.2 In the regime of highintensity driving field, the quantum kinetic equations were solved numerically in the timedomain with initial and boundary conditions. The initial conditions reflect the fact that at ambient temperature the electrons in the CNT are distributed according to the Fermi equilibrium distribution. The boundary conditions are the conditions of the periodicity of the solution in the quasimomentum space. We considered the interaction of the CNT with the femtosecond laser pulse in the vicinity of the plasma resonance. The spectra of the induced current density were obtained. All calculations were performed for the metallic zigzag (9,0) CNT. Practically complete suppression of the highorder harmonic generation near this resonance was observed. The temporal behavior of the induced current was investigated in the case of exact resonance. Quasiperiodic relaxation of the induced current was revealed with time intervals much longer than the input pulse. Similar behavior was observed in metallic clusters. Thus, it is possible to conclude, that the carrier dynamics in CNTs, exposed to electromagnetic pulses, exhibits properties common with those of plasma. The research was partially supported by INTAS and the State Committee for the Science and Technology of Belarus under project 03504409, and by the NATO Science for Peace Program under project SfP981051. A. M. N. is grateful to the World Federation of Scientists for the financial support.
References 1. G. Ya. Slepyan, A. A. Khrutchinskii, A. M. Nemilentsau, S. A. Maksimenko, and J. Herrmann, Highorder optical harmonic generation on carbon nanotubes: quantummechanical approach, Intern. J. Nanoscience 3, 343354 (2004). 2. X. Liu, J. Si, B. Chang, G. Xu, Q. Yang, Z. Pan, S. Xie, P. Ye, J. Fan, and M. Wan, Thirdorder optical nonlinearity of the carbon nanotubes, Appl. Phys. Lett. 74, 164166 (1999).
HYDRODYNAMIC MODELING OF FAST ION INTERACTIONS WITH CARBON NANOTUBES DUNCAN J. MOWBRAY, SANGWOO CHUNG, ZORAN L. MIŠKOVIû Department of Applied Mathematics, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1
Abstract. We use the twofluid, twodimensional hydrodynamic model, proposed in (Mowbray et al., 2004), to describe the collective response of a carbon nanotube's ı and ʌ electrons to fast ions moving parallel to the nanotube. In the present contribution, two extensions of this model are being made: (1) to the case of multiwalled carbon nanotubes, where the interwall electrostatic interaction gives rise to rich plasma spectra, and (2) generalization to include a dielectric medium surrounding the nanotube. Keywords: hydrodynamic modeling; carbon nanotubes; ion channeling
The twofluid model (Mowbray et al., 2004) treats the ʌ and ı electrons of a single wall nanotube (SWNT) as separate twodimensional fluids confined to the same cylindrical surface, while the electrostatic interaction between these fluids gives rise to a splitting of the plasmon energies in qualitative agreement with experiments.
Figure 1. A MWCNT with ten walls of inner radius ain = 3.6 ǖ and interwall separation ǻ = 3.4 ǖ exhibits dispersion curves (in eV) for the ı + ʌ and ʌ plasmons, shown versus longitudinal momentum k (in ǖ1) for (left) m = 0 and (right) m = 1. 177 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 177–178. © 2006 Springer. Printed in the Netherlands.
178 The plasmon dispersion curves for a MWCNT with a number of walls N = 10 are shown in Fig. 1. The twofluid model reproduces the ı + ʌ bands found by Yannouleas et al. in (Yannouleas et al., 1996), as well as the quasiacoustic ʌ bands. Note that the ten ʌ bands from each wall coincide for m = 0 as shown in Fig. 1, left, while splitting of the acoustic bands occurs for higher m values as shown in Fig. 1, right, for m = 1.
Figure 2. Stopping Power (in atomic units) for an ion moving with speed v along the nanotube axis is shown for (left) a MWCNT with ain = 3.6 ǖ and interwall separation ǻ = 3.4 ǖ in vacuum for N = 1 (dashed line), 2 (dotted line), 10 (thin line), and 20 (thick line) walls and (right) a SWCNT with ain = 7 ǖ in vacuum (thin line) and embedded in a dielectric media of İ = 4 (thick line). The lowvelocity peaks correspond to energyloss to the quasiacoustic plasmons.
The twofluid model stopping power for N = 1, 2, 10, and 20 walls is shown in Fig. 2, left, for an ion traversing the carbon nanotube axis. We notice that as the number of walls increases, the stopping power becomes much more pronounced at higher velocities. In Fig. 2, right, we find that an embedding dielectric media of İ = 4 causes the stopping power of a SWCNT to both decrease and shift to lower velocities, as found for metallic nanotubes (Prodan et al., 2002). These results suggest that using SWCNTs and an embedding dielectric media may be employed to reduce the stopping power in the high velocity regime and facilitate channeling of fast ions. References Mowbray, D. J., Miškoviü, Z. L., Goodman, F. O., and Wang, Y. N., 2004, Interactions of fast ions with carbon nanotubes: twofluid model, Phys. Rev. B 70:195418/17. Prodan, E., Nordlander, P., and Halas, N. J., 2002, Effects of dielectric screening on the optical properties of metallic nanoshells, Chem. Phys. Lett. 368:94101. Yannouleas, C., Bogachek, E. N., and Landman, U., 1996, Collective excitations of multishell carbon microstructures: Multishell fullerenes and coaxial nanotubes, Phys. Rev. B 53(15):1025510259.
LOCAL RESISTANCE OF SINGLEWALLED CARBON NANOTUBES AS MEASURED BY SCANNING PROBE TECHNIQUES BRETT GOLDSMITH, PHILIP G. COLLINS* Department of Physics and Astronomy, University of California at Irvine, Irvine CA 926974576
Abstract. Singlewalled carbon nanotube (SWCNT) resistance arising from point defects can be directly imaged using scanning probe techniques. Here, we combine Scanning Gate Microscopy (SGM) and Kelvin Force Microscopy (KFM) to electronically identify defect sites and study their contributions to SWCNT resistances.
Keywords: carbon nanotube; Kelvin gate microscopy; scanning gate microscopy; defect; local resistance
A ballistic SWCNT is theoretically predicted to have a 6.5 k: resistance based on the presence of two onedimensional channels for conduction. In real devices, however, the experimental resistance is typically higher than this measured value. Various mechanisms contribute to this increased resistance. In long SWCNTs, diffusive scattering plays a significant role. In shorter SWCNTs, contact effects can be dominant. In short devices with ohmic contacts, however, a significant number of devices continue to exhibit unexpectedly large resistances. This variability poses a major problem for both the fundamental understanding and commercial application of SWCNT circuits. Scanning probe techniques provide a novel way to directly image the electric potential gradients along a SWCNT and identify the sources of resistance.1,2 In this study, the contact resistance was minimized through proper processing of the SWCNT circuit and devices were selected which nevertheless had large, gatedependent resistances. Nearly 10% of the devices from a waferscale fabrication batch met these criteria. Figure 1, left, shows an example of two voltage profiles acquired along a particular SWCNT device using KFM. At different sourcedrain biases, the SWCNT exhibits localized voltage drops at particular points along its length. It *To whom correspondence should be addressed. Philip G. Collins; email: [email protected]
179 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 179–180. © 2006 Springer. Printed in the Netherlands.
180 is immediately apparent that the SWCNT does not smoothly ramp from 2 to 0V as for a diffusive conductor. Instead, sudden jumps and changes in slope are observed. Furthermore, KFM cannot identify a difference between the surface potential of the electrode and that of the SWCNT near the contact, indicating that the majority of the resistance is not due to contact effects. Instead, all of the surface potential gradients occur along the SWCNT body.
Figure 1. Left: KFM voltage profile along a SWCNT biased at 2.0V (top) and at 3.0V (bottom). Both profiles exhibit sharp voltage drops identifying locally resistive sites. Source and drain electrodes connect the SWCNT near the edges of both plots. Right: A SWCNT imaged topographically by AFM (top) and electronically by SGM (middle). A linescan parallel to the SWCNT identifies local regions dominating the transconductance (bottom).
The same SWCNT can also be characterized by SGM, in which the AFM tip is used as a local gate as shown in Fig. 1, right. The lack of gatesensitive Schottky barriers in this figure confirms that this sample is likely to be metallic, despite twoterminal behavior of a fieldeffect transistor. The SGM measurement identifies nearly all of the gatedependence as being due to local points of resistance along the SWCNT body. These effects are consistent with what we might expect form a defects in or near the SWCNT wall. Further analysis indicates that the sites of gate sensitivity in Fig. 1, right, have a onetoone correspondence with the sites of local resistance identified in Fig. 1, left. The primary conclusion is that defects in SWCNTs are statistically relevant and, when present, they can dominate the electronic characteristics of a device, even causing metallic SWCNTs to behave like semiconducting ones. References 1. M. Freitag, A. T. Johnson, S. V. Kalinin, and D. A. Bonnell, Role of single defects in electronic transport through carbon nanotube fieldeffect transistors, Phys. Rev. Lett. 89, 216801 (2002). 2. A. Bachtold, M. S. Fuhrer, S. Plyasunov, M. Forero, E. K. Anderson, A. Zettl, and P. L. McEuen, Scanned probe microscopy of electronic transport in carbon nanotubes. Phys. Rev. Lett. 84(26), 60826085 (2000).
BAND STRUCTURE OF CARBON NANOTUBES EMBEDDED IN A CRYSTAL MATRIX P. N. D'YACHKOV, D. V. MAKAEV Institute of General and Inorganic Chemistry, Russian Academy of Sciences, Leninskii pr. 31, 119991 Moscow, Russia,
Abstract. The electronic structure of singlewall carbon nanotubes (SWCNTs) embedded in a crystal matrix is calculated in terms of a linear augmented cylindrical wave method (LACW). In the case of armchair nanotubes, the delocalization of the nanotube electrons into the matrix region is responsible for the high energy shift of the ıstates and growth of the electron density of states at the Fermi level. For the semiconducting nanotubes, it causes a decay of the minimum energy gap and a formation of a metallic state.
Keywords: Electronic structure; carbon nanotube; crystal matrix; nanotube metallization effect; linear augmented cylindrical wave method
There have been attempts to integrate SWCNTs in conventional 2D or 3D semiconductor systems. For example, encapsulated singlewalled SWCNTs in epitaxially grown semiconductor heterostructures were found promising for the realization of new semiconductor hybrid devices such as a SWCNT contacted by a twodimensional electron gas.1 Here, we present a simple model for electronic structure of embedded SWCNTs.2 Our approach is based on a LACW method.3 The method is just a reformulation of the wellknown linear augmented plane wave theory for cylindrical multiatomic systems. In the LACW method, the oneelectron potential is constructed using the muffintin approximation. The oneelectron potential is spherically symmetric in the region of the atoms and is constant in the interstitial region up to the two cylindrical potential barriers. In the case of embedded SWCNT, there is no vacuum but a matrix region outside the tubule. The potential barrier Vm between the SWCNT and the matrix is penetrable so that an electronic tunneling exchange between the SWCNT and matrix is possible. For simplicity, we treat the matrix as a homogenous medium with a constant potential neglecting the atomic structure of the matrix and the 181 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 181–182. © 2006 Springer. Printed in the Netherlands.
182 complicated structure of the potential barrier on the interface between the SWCNT and the matrix. It is a model of a SWCNT in contact with an electron gas. Then we solve the oneelectron Schrödinger equation for the orbitals and the electronic energies of the SWCNT in the matrix, the potential Vm being treated as a parameter. According to the LACW calculations, the delocalization of electrons of the armchair nanotube into the matrix region results in a strong band structure perturbation. A high energy shift of the ıstates relative to the ʌbands is the most significant effect of the embedding. As a result, these electrons can take part in the SWCNT charge transfer through the mechanism of electron tunneling into the matrix. The interaction with the matrix results in a growth of the electronic density of states at the Fermi level. The electron tunneling into the matrix does not destroy the metallic character of the armchair nanotube band structure. In the case of semiconducting SWCNT, the minimum energy gap is very sensitive to the matrix effect. The gap decreases with decreasing Vm. Finally, the gap is closed up and the semiconducting behavior breaks down. The metallization of the embedded SWCNTs is consistent with the measured transport properties of the tubesemiconductor devices with carbon nanotubes encapsulated in a semiconductor crystal using the molecular beam epitaxy.1 The presented results from a collection of 20 measured devices displayed no gate voltage dependence of the conductance at room temperature, i.e., the incorporated nanotubes were metallic.1 A Luttinger liquid behavior confirmed that the leak currents through the substrate can be neglected and the electronic transport is dominated by the properties of the embedded metallic SWCNTs.1 This work was supported by RBRF (grant 040332251).
Refrences 1. A. Jensen, J. R. Hauptmann, J. Nygerd, J. Sadowski, and P. E. Lindelof, Hybrid devices from single wall carbon nanotubes epitaxially grown into a semiconductor heterostructures, Nano Lett. 4, 349352 (2004). 2. P. N. D’yachkov and D.V. Makaev, Electronic structure of embedded carbon nanotubes, Phys. Rev. B. 71, 081101(R) (2005). 3. P. N. D'yachkov, Augmented waves for nanomaterials, in Encyclopedia of Nanoscience and Nanotechnology, edited by N. S. Nalwa, (American Scientific Publishers, New York, 2004), v. 1, pp. 191212.
MAGNETOTRANSPORT IN 2D ARRAYS OF SINGLEWALL CARBON NANOTUBES V. K. KSENEVICH* Department of Physics, State University of Belarus, 220050 Minsk, Belarus J. GALIBERT Laboratoire National des Champs Magnétiques Pulsés BP 14245 143, Avenue de Rangueil F31432 TOULOUSE CEDEX 4, France L. FORRO Ecole Polytechnique Federal de Lausanne, DPIGA, PH Ecublens, CH1015, Lausanne, Switzerland V. A. SAMUILOV Department of Materials Science, State University of New York at Stony Brook, N.Y. 117942275, U.S.A.
Abstract. Magnetotransport properties of twodimensional (2D) arrays of singlewall carbon nanotubes (SWCNTs) were investigated. The magnetoresistance (MR) in pulsed magnetic fields and the temperature dependence of the resistance of 2D arrays of SWCNTs were measured in the temperature range 1.8–300 K and in fields up to 9 T. The crossover between metallic and nonmetallic temperature dependence of the resistance was observed at T § 125 K. At low temperature, positive MR was observed with decreasing amplitude with the increase of temperature.
Keywords: SWCNTs arrays; magnetoresitance; weak localization
We report magnetotransport measurements on SWCNT arrays. We used LangmuirBlodgett (LB) technique for selfassembling SWCNTs. Chemical functionalization of the SWCNTs1 was realized in order to reach high solubility of the nanotubes in organic solvents and to obtain a monolayer on the water surface. The monolayers of SWCNTs were transferred to the substrate with in To whom correspondence should be addressed. V.K.Ksenevich, Department of Physics, State University of Belarus, 220050 Minsk, Belarus; email: [email protected]
183 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 183–184. © 2006 Springer. Printed in the Netherlands.
184 plane macroscopic fingershaped electrodes. Minimum on the temperature dependence of the resistance of the SWCNTs arrays, R(T), was observed at T § 125K. At T > 125 K, R(T) was found to have a metallictype behavior (dR/dT > 0). In the lowtemperature range (T ~ 4.250 K), R(T) is an inherent for weak localization power law with an exponent of –0.60. Such type of dependence can be interpreted in terms of heterogeneous model for conduction, where the presence of structural defects and intertube connections gives rise to low temperatures localization effects.2 The conductivity of the system at high temperatures is determined by the charge transport trough metallic nanotubes. Positive MR with decreasing amplitude was observed with the increase of the temperature in the low temperature range, as shown in Fig. 1. The localization effects in the SWCNTs arrays are confirmed by the shape of the MR curves.
Figure 1. Magnetoresistance of the SWCNTs array for magnetic field parallel to the surface.
This work was supported by the INTAS grant #03504409.
References 1. 2.
P.W. Chiu, G.S. Duesberg, U. DettlaffWeglikowska, S. Roth. Interconnection of carbon nanotubes by chemical functionalization, Appl.Phys.Lett., 80, 38113813 (2002). A.B. Kaiser, G. Duesberg, S. Roth. Heterogeneous model for conduction in carbon naotubes, Phys. Rev. B 57, 14181421 (1998).
COMPUTER MODELING OF THE DIFFERENTIAL CONDUCTANCE OF SYMMETRY CONNECTED ARMCHAIRZIGZAG HETEROJUNCTIONS O. V. OGLOBLYA Kyiv National Shevchenko University, Kyiv 01033, Ukraine G. M. KUZNETSOVA Scientific Center of Radiation Medicine, Kyiv 04050, Ukraine
Abstract. Computer modeling of the differential conductance of symmetry connected armchair (n,n) and zigzag (2n,0) nanotubes was carried out in the framework of the surface Green's function matching (SGFM) method. It is shown that the heterojunctions (n,n)/(2n,0) have a band gap in the electron energy spectrum. It is demonstrated that the IV curve increases with constant values with the increase of the voltage. The height of the IV curve plateau is bigger for heterojunctions of larger diameter. It is shown that the maximum of the differential conductance of heterojunctions with bigger diameter is at comparatively low voltage. Keywords: differential conductance; nanotube; heterojunction
In order to calculate the conductance of a zigzag/armchair nanotube and a symmetryconnected (n,n)/(2n,0) heterojunction, we use the Kubo formula (Datta, 1995). The conductance is related to the current via I = GV and is given by the Landauer formula G (2e2 / h)T . Here T is the transmission function expressed as T Tr[* R G * L G ] , where G+/ are the retarded and advanced Green's functions, and īR,L describe the coupling of the conductor to the leads. We used the recurrence formulas derived for layered systems (Sancho et al., 1985) in order to calculate the retarded and advanced Green's functions G+/ and self energies Ȉ+/ of isolated nanotubes. Knowing ȈL (selfenergy function of a left semiinfinite tube) and ȈR (same but for a right semiinfinite tube) we can find the coupling functions īL and īR: *{L , R} i[6{L , R} 6{L, R} ] . We considered the symmetry connection of armchair (n,n) and zigzag (2n,0) nanotubes. The nanotubes are positioned on the same axes and connected by n 185 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 185–186. © 2006 Springer. Printed in the Netherlands.
186 topological defects (5/7). In order to calculate the conductance of such heterojunction, we used the SGFM method based on the formula 1 H LR § GL G · § H H L 6 L · ¨ ¸ ¨ H H R 6 R ¹¸ H LR © G GR ¹ © In this formula, H{LR} is the hopping matrix between the two nanotubes, HL and HR are the Hamiltonians of the unit cell of armchair and zigzag nanotubes. ȈL and ȈR are calculated using published formulas (Sancho et al., 1985). The formula above enables us to find G+/ necessary for the calculation of the conductance by the Kubo formula. The results of our calculation for a (4,4)/(8,0) heterojunction are shown on Fig.1, left. We calculated the differential conductance wI / wV assuming phasecoherent transmission and a constant potential within the whole sample (Liang et al., 2001).
Figure 1. Left: density of states (top) and conductivity in units of 2e2/h (bottom) for the (4,4) (dotted line) and (8,0) (dashed line) nanotubes and for the heterojunction (4,4)/(8,0) (solid line). Right: Differential conductance for (a) (4,4)/(8,0), (b) (5,5)/(10,0), (c) (6,6)/(12,0), (d) (7,7)/(14,0), (e) (8,8)/(16,0), (f) (9,9)/(18,0).
The results of the calculation of the differential conductance of (n,n)/(2n,0) heterojunctions of different diameter are shown in Fig. 1, right. The maximum of the conductance for larger diameter heterojunctions shifts to lower voltage.
References Datta, S., 1995, Electronic Transport in Mesoscopic Systems, Cambridge University Press, Cambridge. Lopez Sancho, M. P., Lopez Sancho, J. M., and Rubio, J., 1985, Highly convergent schemes for the calculation of bulk and surface Green functions, J. Phys. F 15:851858. Liang, W., Bockrath, M., and Borovic, D., 2001, FabryPerot interference in a nanotube electron waveguide, Nature 411:665669.
Part IV. Molecule adsorption, functionalization and chemical properties
MOLECULAR DYNAMICS SIMULATION OF GAS ADSORPTION AND ABSORPTION IN NANOTUBES ANA PROYKOVA* University of Sofia, Faculty of Physics, 5 J. Bourchier Blvd., Sofia1126, Bulgaria
Abstract. Physi and chemisorption are considered within the framework of classical and quantum molecular dynamics and density functional theory. We analyze coating (adsorption of metal monolayer) of carbon nanotubes as the cause of gas adsorption enhancement. Coarsegraining of less relevant degrees of freedom to obtain Hamiltonians spanning large length and time scales is a successful approach. The diffusion of adsorbed substances is modeled as well. Carbon nanotubes are regarded as mass conveyors as an application of diffusion. Charge fluctuations of electron density of the adsorbed molecules and of the carbon tube are the sources of metastability of the physisorption states.
Keywords: adsorption; simulations of carbon nanotubes; Molecular Dynamics method; coarsegraining technique; abinitio calculations; timescales
1. Adsorption – general characteristics Adsorption of one or more of the components at one or more of the phase boundaries of a multicomponent, multiphase system is said to occur if the concentrations in the interfacial layers are different from those in the adjoining bulk phases, so that the overall stoichiometry of the system deviates from that corresponding to a reference system of homogeneous bulk phases whose volumes and/or amounts are defined by suitably chosen dividing surfaces or by a suitable algebraic method.
*To whom correspondence should be addressed. Ana Proykova, University of Sofia, Faculty of Physics, 5 J. Bourchier Blvd., Sofia1126, Bulgaria; email: [email protected]
187 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 187–207. © 2006 Springer. Printed in the Netherlands.
188 1.1. PHYSISORPTION AND CHEMISORPTION
A qualitative distinction is usually made between chemisorption and physisorption. The problem of distinguishing between chemisorption and physisorption is basically the same as that of distinguishing between chemical and physical interaction in general. No absolutely sharp distinction can be made and intermediate cases exist, for example, adsorption involving strong hydrogen bonds or weak charge transfer. In terms of the relative binding strengths and mechanisms, a strong 'chemical bond' is formed between the adsorbate atom or molecule and the substrate. In the case of chemisorption the adsorption energy, Ea, of the adatom is of a few eV/atom. Physisorption (or physical adsorption) results from the presence of van der Waals attractive forces due to fluctuating dipole (and higher order) moments on the interacting adsorbate and substrate with no charge transfer or electrons shared between atoms. These intermolecular forces, usually between closedshell systems, are of the same kind as those responsible for the imperfection of real gases and the condensation of vapours. Table 1. The following features characterize physisorption and chemisorption. Physisorption
Chemisorption
Evidence for the perturbation of the electronic states of adsorbent and adsorbate is minimal The chemical nature of the adsorptive is not altered by adsorption and subsequent desorption
Changes in the electronic state
Energies are of order 50500 meV/atom
The energy is of the same order of magnitude as the energy change in a chemical reaction between a solid and a fluid, eV/atom
The elementary step in physisorption from a gas phase does not involve activation energy
The elementary step in chemisorption often involves activation energyi (adsorption sites)
multilayer adsorption or filling of micropores
one layer of chemisorbed molecules is formed
The chemical nature of the adsorptive(s) may be altered, i.e. the chemisorption may not be reversible.
Physisorption energies are of order 50500 meV/atom. As they are small, they can be expressed in K/atom, via 1 eV { 11604K, omitting Boltzmann’s constant in the corresponding equations. One can see that these energies are comparable to the sublimation energies of rare gas solids. The physical adsorption is believed to serve as a precursor, which enhances the transition to the chemisorption state. A first, precursor stage, has all the characteristics of physisorption, but this state is metastable. In this state the molecule may reevaporate, or it may stay on the surface long enough to transform irreversibly into a chemisorbed state.
189 This transition from physisorption to chemisorption usually results in a split of the molecule and adsorption of individual atoms: dissociative chemisorption. If the heat of adsorption is given up suddenly and is imparted to the resulting adatoms, then the dissociation stage is explosive as it is in the case of F2. The adsorption energies for the precursor phase are similar to phyisorption of rare gases, but may contain additional contributions from the dipole, quadrupole, or higher molecular multipoles. 1.2. LOCALIZED ADSORPTION AT LOW COVERAGE
Equilibrium phenomena are described by thermodynamics, which is the dynamics of heat or by statistical mechanics, which works well for large (on the order of Avogadro’s number) numbers of particles – atoms and /or molecules. However, many properties of nanosized (small number of particles, away from the thermodynamics limit) materials are concerned with kinetics, where the rate of change of metastable structures (or their inability to change) is dominant (Proykova et al., 2001; Proykova, 2002). An equilibrium effect is the vapor pressure of a crystal (pure element). A kinetic effect is crystal growth from the vapor. The surface coverage for both monolayer and multilayer adsorption is defined as the ratio of the amount of adsorbed substance to the monolayer capacity. For chemisorption the monolayer capacity is defined as the amount of adsorbate, needed to occupy all adsorption sites as determined by the adsorbent structure and by the chemical nature of the adsorptive; for physisorption it is the amount needed to cover the surface with a complete monolayer of molecules in a closepacked array (Venables, 2000). The sublimation of a pure solid at equilibrium is given by the condition Pv = Ps, where Pv is the chemical potential of the vapor and Ps is the chemical potential of the solid. At low pressure p, P v is:
Pv
kT ln(kT / pO 3 ),
(1)
1/2
where O = h/(2SmkT) is the thermal de Broglie wavelength; h is the Planck constant; k is the Boltzmann constant; m is the electron mass. The equilibrium vapor pressure peq in terms of the chemical potential of the solid is
peq
(2S m / h 2 )3/ 2 (kT )5 / 2 exp( P s / kT ).
(2)
To calculate the vapor pressure we usually model Ps at a low pressure assuming harmonic vibrations of the solid at a given lattice constant. The free energy per particle is
190 F/N
P s U 0 3h / 2 ! 3kT ln(1 exp(h / kT )) !,
(3)
where < > denote averaged values. The (positive) sublimation energy at zero temperature is L0
(U 0 3hQ / 2 !),
(4)
where the first term is the (negative) energy per particle in the solid relative to vapor, and the second is the (positive) energy due to zeropoint vibrations. The vapor pressure is important at high temperatures, where the Einstein model (all 3NQ's are the same) of the solid becomes realistic if thermal expansion is taken into account in U0. This model gives
ln(1 exp(h / kT )) ! ln(h / kT ) !,
(5)
3
so that exp(Ps/kT) = (hQ/kT) exp(L0/kT) and
peq
(2S mQ 2 )3/ 2 (kT )1/ 2 exp( L0 / kT ).
(6)
1/2
Thus peqT follows the Arrhenius law and the preexponential factor depends 3 on the lattice vibration frequency as Q . The missing Planck’s constant h shows 1/2 a classical effect with equipartitioning of energy. Since the T term slowly varies most tabulations of vapor pressure give the constants A and B from log10peq = A  B/T. The values for L0 and Q are obtained along these lines. The chemical potential Pa of the adsorbed layer satisfies two possible conditions: Pa = Pv for equilibrium with the vapor or Pa = Ps for equilibrium with the solid. The canonical partition function for the adsorbed atoms is Za = ¦iexp(Ei/kT) with F = kTlnZ. For Na adsorbed atoms distributed over N0 sites with the same adsorption energy Ea, Za = Q(Na,N0)exp(NaEa/kT), where Q is the configurational (and vibrational) degeneracy. The configurational entropy appears since, at low coverage, there are many ways to arrange the adatoms on the available adsorption sites, (Hill 1987): Q = N0!/(Na!)(N0  Na)! multiplied by qNa if vibrational effects are included. The expression for lnQ is evaluated using Stirling's approximation for N! = NlnN  N to give Pa = F/Na = (kT/Na)lnZa = kTln(T/(1  T))  Ea  kT lnq and T = Na/N0. The first term is the configurational contribution in terms of the adatom coverage T, the second term is the adsorption energy (measured positive with the vacuum level zero), and the last term is the (optional) vibrational contribution. The density of adatoms is determined by Ps = 3kTln(hQ/kT)  L0 = Pa in the high temperature Einstein model. Thus at low coverage, T is
T
Na / N0
C exp[( L0 Ea ) / kT ],
(7)
191 where the preexponential function depends on vibrations in the solid, and the important exponential term depends on the difference between the sublimation and the adsorption energy. If C is unknown, we usually put it equal to 1 to obtain a first estimate of T at a given temperature. The Langmuir Adsorption Isotherm results from Pa = Pv. Using this to calculate the vapor pressure p in equilibrium with the adsorbed layer we get p C1T /(1 T ) exp( Ea / kT ), 1
(8) 3
or T = [(T)p/[1  [(T)p], with [(T) = C1 exp(Ea/kT); the constant C1 = kT/qO . The coverage is linearly proportional to p for small p approaching 1 as p o f. The other limit of isolated adatom behaviour is the 2D gas (Venables 2000). This case is appropriate to a very smooth substrate, with shallow potential wells. The mobile adatoms see the average adsorption energy E0, and gain additional entropy from the gaseous motion. The chemical potential is now Pa = E0 + kTln(NaO2/Aqz), where this expression is valid at sufficiently low density for the distinction between classical BoseEinstein and FermiDirac statistics not to be important. The derivation involves evaluating the partition function by summing over 2D momenta, analogously to a 3D gas, while retaining the zmotion partition function qz. The 3D to 2D difference accounts for O2 rather than O3, and the Na/A, the number of adsorbed atoms per unit area, is the 2D version of the 3D density N/V, as in pV = NkT. The perfect gas law has the 2D form )A = NakT, where ) is the spreading pressure (Venables 2000). This means that Pa = E0 + kTln()O2/kT), or Pa = E0 + P2 + kTln), where P2 = kTln(kTqz/O2) is the standard free energy for the 2D gas. This makes the correspondence between 3D gases and 2D adsorption clear: p l ), P0 l P2; the energy is lower in the 2D case by E0. The various q's are dimensionless and often skipped. Henry's Law for 2D gas adsorption is obtained by putting Pa = Pv: p C2 ( N a / A) exp( E0 / kT ),
(9)
or (Na/A) = F'(T)p with F'(T) =C21exp (E0/kT) and C2 = kT/qzO. The 2D gas form has (Na/A) proportional to p, whereas the localized form has the coverage T = Na/N0 proportional to p. These could be reconciled if we write (Na/A) = T(N0/A). The monolayer coverage (N0/A) and the area density of adsorbed atoms (Na/A) have been defined as constants. Both N0 and Na are numbers, not area densities, although they are densities for A = 1. One can rewrite the 2D gas equation as p (kTN 0 / Aqz O ) exp( E0 / kT ),
(10)
which is in a form that can be compared directly with the corresponding equation for localized adsorption. This comparison shows that there is a
192 transition from localized to 2D gaslike behaviour as T is raised, because Ea > E0, whereas the preexponential (entropic) term is larger for the 2D gas. The ratio of coverages at a given p for the two states is
T gas / Tloc
(2S mkT / h 2 qx q y )( A / N 0 ) exp[( E0  Ea ) / kT ] (2S ma 2Q d2 / kT ) exp[( E0  Ea ) / kT ],
(11)
where the length a is an atomic dimension (a2 = A/N0), and the last equality is only true for the high temperature limit where equipartition of energy holds (no term in h). We can state that the localized atoms will vibrate with amplitude and 4S2m2Qd2 is the energy associated with this 2D oscillation, which equals 2kT at high T, assuming a harmonic approximation is valid. Thus, the preexponential is just a ratio of free areas (a2/S2), the numerator associated with the 2D gas, and the denominator with the potential well in which the adatom vibrates. This is the basis of cell models of lattice vibrations, introduced originally by LennardJones and Devonshire in 1937 (Hill, 1986). The free area is defined by integrating the Boltzmann factor over the 'cell' in which the atom vibrates; in 3D this produces a free volume. This approximate classical theory is very effective in computations, since it includes thermal expansion (the response to the spreading pressure exerted by the vibrations), which more sophisticated models ignore. It also does not rely on a harmonic approximation; for these reasons at least, it deserves to be better known. 1.3. SPECIFICITY OF ADSORPTION ON SWCNT AND SWCNT BUNDLES
One of the most interesting issues in adsorption is the basic question of the system dimensionality. A specified nanotube adsorption problem can exhibit behavior characteristic of one, two or three dimensions depending on thermodynamic (temperature and number of particles) or microscopic (atomic size relative to nanotube radius) parameters, and geometry (independent tubes or bundles). The study of adsorption of atoms on nanotube surfaces is also essential for nanoelectronics aiming to achieve low resistance ohmic contacts to nanotubes and to fabricate functional nanodevises. It is important for nanotechnology to produce nanowires with controllable sizes. While the adsorption on the surface of a SWCNT is a twodimensional process, the diffusion of substances in the tubes is onedimensional process. 1.3.1. Adsorption on SWCNT It was shown with the help of firstprinciples gradientcorrected densityfunctional calculations (Pati et al., 2002) that the adsorption of water molecules
193 on a SWCNT causes a reduction of the electronic conduction due to charge transfer between the adsorbate and the SWCNT: 0.03e is transferred from a single water molecule to the nanotube. The authors conclude a physisorption although they claim a weak bond between a hydrogen atom of a water molecule and one C atom. A systematic study of single atoms adsorption on a carbon nanotube has been performed by Durgun et al., 2003. Higher coverage and decoration of adsorbed foreign atoms can produce nanostructures such as nanomagnets, nanometer size magnetic domains, onedimensional conductors and thin metallic connects, which could be used in technological applications as spintronics and high density data storage. The dorbitals of the transition metal atoms are responsible for a relatively high binding energy that displays a variation with the number of filled dstates. The main result of the experimental study of Zhang et al. (2000, 2000a) that continuous nanowires of any metal can be obtained by coating the SWCNT with titanium. The strong interaction between Ti and SWCNT can be understood in the frame of density functional theory, which will be described in some detail in the Section 2.1. The average exchange time between methane molecules adsorbed inside SWCNTs and the free gas outside is estimated to be about 80 ms from nuclear magnetic resonance measurements (Kleinhammes et al., 2003). 1.3.2. Adsorption in bundles of SWCNT Experimental results of methane adsorption on closedended singlewall nanotube bundles (Talapatra, 2002) give the isosteric heat of adsorption as a function of the amount of methane adsorbed on the SWNT substrate for coverage in the first layer. The isosteric heat of adsorption is found to decrease with increasing coverage. This behavior provides an explanation for differences in previously reported values for this quantity. Substantial agreement was found with previous reports on the temperature dependence of the pressures at which the two firstlayer substeps occur for methane adsorbed on SWNT bundles. Grand canonical MonteCarlo simulations with a LennardJones potential have been performed to study the adsorption of methane, X and Ar onto bundles of closed SWCNT (Shi, 2003). The results agree with the experiments of Talapatra, et al., 2002: four different adsorption sites have been identified on bundles of SWCNTs – internal (endohedral), interstitial channels (ICs), external groove sites, and external surfaces. The main conclusion is that the ICs become important in the case of heterogeneous bundles. However, there is no obvious plateau region corresponding to groove site filling for the heterogeneous bundles for methane and Ar and the authors conclude that the groove sites for both homogeneous and heterogeneous bundles are very similar.
194 The kinetics of adsorption and desorption of methane, Xe, SF6, and other weakly bonded gases from SWCNT bundles has been compared with that from graphite by Hertel (2001). The observed trends in the binding energies of gases with different van der Waals radii suggest that the socalled groove sites on the external bundle surface are the preferred low coverage adsorption sites due to their higher binding energy. The measured sticking coefficients can be related to the diffusion kinetick of adsorbates into the bulk of the nanotube samples. 1.3.3. Adsorption of methane on graphene Methane is of great interest both as a greenhouse gas and as a possible source of hydrogen for fuel cells. A lot of articles have been published on the properties of methane. Our interest in methane has been rooted in the ability of methane to form clathrates or gas hydrates in very cold water under a high pressure. We have revisited the problem of hydrophobicmolecule influence on liquid and solid water structures (Daykova, 2002) in order to understand the competition between the enthalpy and the entropy of clathrate formation. We have suggested a statistical model of the rearrangement of water molecules when methane molecules are introduced in the ideal tetrahedral hydrogenbond network. The most important basis of that model, which considered the methane molecules as rigid objects, has been the hydrophobic interaction. Methane molecules are less rigid if they interact with carbon atoms. The electronic structure of graphene, a single planar sheet of sp2bonded carbon atoms, provides the basis for understanding the electronic structure of singlewalled carbon nanotubes. Recently a lot of efforts have been put into developing applications of SWCNT as gas sensors as the electronic and transport properties of the SWCNTs are significantly changed if they are exposed to gases (Collins 2000). The gas molecules are either physisorbed or chemisorbed depending on the reactivity of the gas. A microscopic theory of physisorption applied for the noble He atoms on variety of free electron metals showed a little influence of the boundary conditions for a nonlocal model on the equilibrium distance (Landman, 1975). This justifies the study of adsorption of methane (resembling the noble gases in interaction with other atoms) on an isolated piece of graphene sheet. Extensive timedependent densityfunctional calculations (DFT) have been performed (Daykova, 2005) with the realspace DFT code (Chelikowsky, 2000) to study the mechanism of adsorption of methane and to check the influence of the slab finite size on the dynamics of the process and the electron distribution as a result of open "ends" of the slab. Thus, no boundary conditions have been introduced. The computations without periodic boundary conditions are more demanding. Hence, a preliminary research for the volume and its shape should be done. In the present study a cubic box with a sidelength of 17.46 ǖ has been
195 chosen for the graphene slab with a diameter of 8.98 ǖ, Figure 1 (a). The grid of calculations is 110 x 110 x 110 with a spacestep of 0.3 Bohr = 0.1587 ǖ; the cutoff radius is 5.29 ǖ; the time step of electron density integration is 1.2 fs; constanttemperature molecular dynamics (MD) calculations at each time step with a frictionmass of 0.0004 a.u. is used for optimizing the configuration at T = 0 K. Under these conditions the electron density at the cutoff radius reduces to 5.4 109 with a respect to the maximum density. There is no need for increasing the box size or the cutoff radius. Details of procedure have been published by the code authors (Vassiliev et al., 2002) and (Proykova, 2003). The center of mass of methane molecule is above the point 0 on the graphene sheet, Fig. 1 (a), initially at a height of 2.91 ǖ. The height is optimized with a combined DFT and MD calculations to be 3.07 ǖ. We have considered two different orientations of the methane molecule relative to the graphene surface. Case 'F', Fig. 1 (b), is characterized with a binding energy of 120 ± 10 meV, which shows a physisorption and agrees surprisingly well with previous calculations of Muris et al. (2001) based on a constanttemperature Monte Carlo method. They have found that the binding energy for a single methane molecule is –121 ± 10 meV for the curved graphitic surface; for planar graphite it is –143 ± 10 meV. The case 'V' denotes a methane molecule rotated at 180q with respect to the position in Fig, 1 (b). In this case, one CH bond is closer to the graphene slab and the other three are further. The binding energy is larger manifesting a configuration dependence reported for Li adsorbed on CNT (Udomvech, 2005). The methanegraphene interaction induces a local curvature of the graphene: in the 'F' case it is hardly seen, Figure 1 (b), while in the 'V' case it is essential. The oscillations of CH4 center of mass in Fig, 1 (c) reveal a faster dynamics in the 'F' case. This result combined with the smaller adsorption energy in the 'F' case, shows that this case is less stable.
196
Figure 1. The CH4 molecule is initially at r0 = 2.91 ǖ above point 0 in the graphene sheet (a); 1, 2, 3, and 4 label the bonds between C atoms in the sheet and appear in the figures (e) and (f); (b) the 'F' orientation of CH4; (c) the relative amplitude of induced oscillations of the CH4 center of mass for the cases 'F' and 'V'; (d) the kinetic energy of the system (molecule and graphene) as a function of time: after 150 fs the temperature is almost 0 K; the oscillations of graphene bonds '1', '2', '3', and '4' for 300 fs: (e) 'V' case; (f) 'F' case.
197 To find out the mechanism of adsorption we study the electron distribution change in both methane and graphene slab due to methanegraphene interaction. Additionally, changes of the fundamental modes of methane bring information about the interaction. We compare our calculations with the experimentally determined values for the fundamental modes of methane obtained by Allan (2005) with the help of electron energy loss spectroscopy, Table 2. Table 2. Vibrational modes of methane (symmetry T ).d Mode
Symmetry species
Type
Frequency (meV)
Line width (meV)
Activity
Q1 Q2 Q3
A1 E T2
Symmetric stretch Twisting Asymm. stretch
361.7 190.2 374.3
30
Raman Raman IR
Q4
T2
Scissoring
161.9
16
IR
Note that the outer graphene bonds (lines 3 and 4 in Fig. 1, (e) and (f)) oscillate with a period of 11 fs, which is the same as the asymmetric stretch mode 11.05 fs (374.3 meV) of the methane molecule. Summation of the two wave processes cause beating which we see in the dynamics of the electron distribution in the methane molecule. 1.4. RELATIONSHIP OF POTENTIAL ENERGY CURVES TO ELECTRONIC SPECTRA OF A DIATOMIC MOLECULE ADSORBED ON A SURFACE
To understand differences in adsorption of various atoms/molecules we inspect the potential energy (Morse) curve, Fig. 2 (a), for a diatomic molecule versus the internuclear distance. The minimum and maximum values for the bond distance for the vibrational state are at A and B (see caption to the figure). At those points the atoms change direction, so the Vibrational Kinetic energy is zero. At re (equilibrium internuclear distance) the Vibrational Kinetic energy has a maximum value and the Vibrational Potential energy is zero. Each horizontal line represents a vibrational state. The ground state is v0 and thee excited states are v1, v2, v3, etc. Each excited state contains a series of different vibrational energy levels and may be represented by a Morse curve. Each vibrational level vn is described by a vibrational wave function Ȍvib. The square of wave function gives the probability distribution, and in this case Ȍvib2 indicates the probable internuclear distance for a particular vibrational state (the dotted line). The higher the line the more probable the internuclear distance is. The most probable distance for a molecule in the ground state is re, while there
198 are two probable distances that correspond to the two maxima in the next vibrational state; three in the third, Fig. 2 (b). In the excited vibrational levels for the ground state and the excited electronic states, there is a high probability of the molecule having an internuclear distance at the ends of the potential function.
Figure 2. (a) Morse curve: Along the AB line we see that the energy is constant while the bond distance (b) is changing, that means the molecule is vibrating, and AB is a vibrational state.
2. Timescales in Molecular Dynamics. Abinitio (quantum) Molecular Dynamics and coarse graining techniques To determine properties of materials one needs to approximate both atomic and electronic interactions. However, these interactions have completely different time and space scales. Because of the large difference between the masses of electrons and nuclei and the fact that the forces on the particles are the same, the electrons respond instantaneously to the motion of the nuclei. Thus, the
< (R I , rn ) < el ({ rn };{R I }) < l ({R I }) nuclei are adiabatically treated and separated from the electrons. The manybody wave function is factorized within the BornOppenheimer approximation.
199 The adiabatic principle reduces the manybody problem in the solution of the dynamics of the electrons in a frozenin configuration {RI} of the nuclei. This is an example of how some degrees of freedom are eliminated in the multiscale modeling. The structure of matter is continuous when viewed at large length scales and discrete at an atomic scale. Modeling techniques are being developed to bridge the gap between the lengthscale extremes. Nanotechnology challenges the researchers to develop and design nanometer to micrometersized devices for applications in new generations of computers, electronics, photonics and drug delivery systems. Theory and modeling play more and more important role in these areas to reduce costs and increase predictability for the new properties. Multiscale materials modeling approaches are required to combine the atomistic and continuum scale. A detailed overview of recent achievements has been completed by Ghoniem and collaborators (Ghoniem et al., 2003). The SWCNT is a good example of the dual behavior of the matter. Their functionalization is done at an atomic level (bonding manipulation) while their elastic modulus shows up at the continuous level. 2.1. DENSITY FUNCTIONAL AND MOLECULAR DYNAMICS METHOD
The CarParrinello (CP) method was applied to perform both dynamic and static calculations. The method performs the classical MD, which computes a time dependent trajectory of all the atomic motions by numerically integrating equation of motion, and simultaneously applies DFT to describe the electronic structure, using an extended Lagrangian formulation. The characteristic feature of the CarParrinello approach is that the electronic wave function, i.e. the coefficients of the plane wave basis set, are dynamically optimized to be consistent with the changing positions of the atomic nuclei. The actual implementation involves the numerical integration of the equations of motion of secondorder Newtonian dynamics. A crucial parameter in this scheme is the fictitious mass associated with the dynamics of the electronic degrees of freedom. Nosé (1984) has modified Newtonian dynamics so as to reproduce both the canonical and the isothermalisobaric probability densities in the phase space of an Nbody system. Nosé did this by scaling time (with s) and distance (with V1/D in D dimensions) through Lagrangian equations of motion. The dynamical equations describe the evolution of these two scaling variables and their two conjugate momenta ps and pv. In practice, the fictitious mass has to be chosen small enough to ensure fast wave function adaptation to the changing nuclear positions on one hand and sufficiently large to have a workable large time step on the other. In our work (Daykov and Proykova, 2001) the equations
200 of motion were usually integrated using the Verlet algorithm (Verlet, 1967). The fictitious mass limits the time step. The CP simulations can be performed using the CPPAW code package developed by Blöchl (Blöchl, 1994). It implements the abinitio (from firstprinciples) molecular dynamics together with the projector augmented wave (PAW) method. The PAW method uses an augmented plane wave basis for the electronic valence wave functions, and, in the current implementation, frozen atomic wave functions for the core states. Thus it is able to produce the correct wave function and densities also close to the nucleus, including the correct nodal structure of the wave functions. The advantages compared to the pseudopotential approach are that transferability problems are largely avoided, that quantities such as hyperfine parameters and electric field gradients are obtained with high accuracy (Petrilli et al., 1998; Blöchl, 2000) and, most important for the present study, that a smaller basis set as compared to traditional normconserving pseudopotentials is required. It is well known that most physical properties of solids are dependent on the valence electrons to a much greater degree than that of the tightly bound core electrons. It is for this reason that the pseudopotential approximation is introduced. This approximation uses this fact to remove the core electrons and the strong nuclear potential and replace them with a weaker pseudopotential which acts on a set of pseudo wavefunctions rather than the true valence wavefunctions. In fact, the pseudopotential can be optimised so that, in practice, it is even weaker than the frozen core potential (Lin et al., 1993). The frozen core approximation was applied for the 1s electrons of C and O, and up to 2p for Cl. For H, C and O, one projector function per angularmomentum quantum number was used for s and pangular momenta. For Cl, two projector functions were used for sand one for pangular momenta. The KohnSham (Kohn and Sham 1965) orbitals of the valence electrons were expanded in plane waves up to a kinetic energy cutoff of 30 Ry. Static DFT calculations can be performed using the atomicorbital based ADF package (Baerends et al., 1973). In these calculations, the KohnSham orbitals were expanded in an uncontracted tripleSlatertype basis set augmented with one 2p and one 3d polarization function for H, 3d and 4f polarization functions for C, O, and Cl. For heavy atoms, relativistic effects become important because electrons near the nuclei move at speeds that are a significant fraction of the speed of light. The electron wave functions near the nuclei must therefore be described by the Dirac equation. However, the pseudopotential method can also be applied here. To determine the radial wave functions, one must work with a generalization of the radial KohnSham equations that correspond to the Dirac equation. The steps in creating a pseudopotential are now modified as follows:
201 Step 1  one must write down and solve a version of the KohnSham equations that generalizes the Dirac equation; Step 2  one draws smooth replacements for the original wiggly wave functions. One can take the original nonrelativistic KohnSham equation and put a pseudopotential in it that produces the wave functions obtained in the previous step. Now the pseudopotential is not only compensating for the smoothing procedure of step 2 that removed wiggles from the wave function, but also is compensating for the differences between the Dirac equation and the Schroedinger equation. That is, the pseudopotential is chosen to produce a wave function for the Schroedinger equation that is also a solution of the original Dirac equation. DFT and MD methods have been simultaneously applied to study conductivity and mechanical properties of (10,0) SWCNT (Iliev et al., 2005). Molecular dynamics simulations have shown that the distribution of the vacancies, not only their concentration, is important for the mechanical strength of the tube. Our calculations demonstrate that the band gap shrinks in a defective tube and the external field changes the band gap differently depending on the axis of application. Insight of carbon nanotube growth has been gained from DFT and tight binding studies (Krasheninnikov et al., 2004). They have found that the adsorption and migration energies strongly depend on the nanotube diameter and chirality, which makes the model of carbon adatom on a flat graphene sheet inappropriate. This result is consistent with our finding (Daykova et al., 2005) that adsorption of methane on a small graphene sheet induces curvature of the sheet, which is not flat anymore. Krasheninnikov and collaborators have used the PAW method to describe the core electrons and the generalized gradient approximation (Perdew and Wang, 1992, Perdew et al., 1992). They have found that adatom migration is a different mechanism from the kickout mechanism as it was previously derived from analytical potential calculations (Maiti, 1967). 2.2. SPACETIME DISCRETE POINTES
For the numerical solution of timedependent wave propagation problems that appear both in DFT and ab initio MD one has to deal with complex geometries with very different evolution scales. The time step of the computations is determined by the fastest process in order to predict reliable results. For a comparison, the electronic degrees of freedom should be integrated in the attosecond domain, while the nuclear degrees of freedom are in the femtosecond domain. If a uniform timestep (fastest process) ǻt is used two problems appear. First, the computational cost is high. Second, the ratio cǻt/(space_step) in the coarser grid will be much smaller than its optimal value. This generates dispersion errors in most numerical schemes. To avoid these problems, it is
202 advisable to keep the ratio constant meaning that ǻt is different in the various space domains determined by the density distribution of the original system. The Figure 3 shows replacement of a uniform timespace mesh with a refined mesh for a complex structure, which is much denser in the middle of the area than at the boarders. At different timesteps (they could be very different) one can use various space discrete steps. The mesh refinement is an opposite approach to coarsegraining techniques, which we will consider in the next section. At the end of the present section we must underline that the use of local time step raises new practical and theoretical problems that are very delicate in the case of hyperbolic equations. There are plenty of articles on the subject, see for example (Manfrin, 1996) and the references therein.
Figure 3. The uniform timespace mesh (left) is replaced by a suitable dynamic mesh (right).
2.3. COARSEGRAINING TECHNIQUE: TIME/TEMPERATURE DEPENDENT PHENOMENA
The study of time/temperature dependent phenomena is more difficult and requires the development of a dynamics for a system with an inhomogeneous level of coarse graining. A nice job has been completed by Stefano Curtarolo in his dissertation for the degree of PhD in material science (Curtarolo 2003). The technique suggested is different from both microscopic (quantum mechanics) and macroscopic (thermodynamics) descriptions. By removing the unnecessary
203 degrees of freedom, the author constructs a local thermodynamics associated with dynamics and interaction between regions. An example useful for the carbon nanotubes is the following. Let us consider a finite one dimensional chain of atoms that interact through nearest neighbor pair potentials. It is possible to remove every second atom by considering each of these atoms as part of a local system with its own local partition function. The local degrees of freedom can be removed by integrating over all the possible configurations of the particular atom. The integration produces a local partition function, local free energy and local entropy, which equals the information that has been locally removed from the system. This construction ‘converts’ the second neighbors into first neighbors with an effective potential which is the microscopic analogy of the pressure for thermodynamic systems. The procedure can be iterated until a multiscale description of the systems is achieved. This approach is rigorously correct at equilibrium being an approximate one to slow dynamics description. Coarse graining requires both a scheme to remove atoms and a recipe for constructing potentials between the remaining atoms. One way is the integration of bond motion similarly to the MigdalKadanoff approach in renormalization group theory (Migdal, 1975; Kadanoff, 1976; Kadanoff, 1977). New potentials can be differently defined satisfying the requirements: a) the coarsegrained system ultimately evolves to the same equilibrium state of the atomistic one; b) the information removed from the original system is quantified by the entropy contribution of each coarse graining. Curtarolo has used as a criterion for the definition of a new potential: the partition function should be unchanged in the coarse graining procedure. 3. Diffusion of the adsorbed gases An important observation has been that adsorbed atoms (adatoms) form small clusters which diffuse on metal surfaces by undergoing possible cluster configurations. The nature of the motion depends on the surface morphology. Examples are onedimensional diffusion of W on the (211) W surface and twodimensional diffusion of Pt on W. What is the dimensionality of gas diffusion on/in nanotubes? Different approaches – experimental, computational, and theoretical – give supplementary information on the specific features of particle diffusion on/in CNT and SWCNT bundles. Theoretically one solves the generic diffusion equation
wc / wt .( Dc )
Q,
(12)
204 where c = Na/A is the concentration of the adatoms, D is the diffusion coefficient and Q is a volume source, may be anisotropic in which case D is a 2x2 matrix. The boundary conditions can be of Dirichlet type, where the concentration on the boundary is specified, or of Neumann type, where the flux n.(Dc) is specified. It is also possible to specify a generalized Neumann condition. It is defined by n.(Dc) + qc = g, where q is a transfer coefficient. This approach does not reflect the discrete nature of SWCNTs, which restricts its applications. Recent research reveals that binary gases and fluids diffuse differently at the nanoscale than many other polyatomic molecules. For example, nbutane and isobutene are predicted to have substantially different separation behaviors when they are mixed with methane molecules in a singlewall carbon tube (Moe, 2001). Further studies to determine the exact mechanisms responsible for this unique trait are, however, a great challenge due to the difficulty of detecting and determining the behavior of the molecules experimentally. Atomistic simulations (Skoulidas et al., 2002) for both self and transport diffusivities of methane and hydrogen in CNT predict orders of magnitude faster transport than that in zeolites. The high rates in CNT result from the inherent smoothness of the nanotubes. The authors used equilibrium dynamics to obtain the diffusion coefficients. There is an interest in Li adsorption and diffusion on/in CNT in relation with Li batteries production. Frontera and collaborators (Garao et al., 2003) studied molecular interaction potential between Li+ and smalldiameter armchair SWCNT based on abinitio calculations. They observed a channel that would allow ion mobility inside the nanotube. Since the Li atom and cation prefer to localize near the carbon nanotube sidewall, smalldiameter carbon nanotubes are more advantageous than the larger ones for batteries usage. The rate of diffusion is inversely related to the binding energy. Density functional computations based on the B3LYP functional and allelectron basis set centered on atoms show that the binding energy depends on the configuration rather than on the diameter size (Udomvech et al., 2005). This result contradicts the finding of Krasheninnikov and thus needs an independent study. To complete this section it is important to note that the diffusion process is the basis of various phenomena including carbon nanotube growth by carbon diffusion; convey of nanoscale materials with the help of SWCNT; migration of defects; gas separation and filtration. One of the future challenges is the development of explicit links between the atomistic scales and the continuum level in all numeric techniques included in the present paper. The issue of computational efficiency must be also
205 addressed so that truly largescale simulations on thousands of processors can be effectively performed. ACKNOWLEDGEMENTS
Computations on the methane adsorption on graphene have been performed with a partial support from HPCEurope (CINECA, Italy). The results for the influence of external electric field on defective SWCNT have been calculated under the TaiwanBulgarian project. A Grant from the Ministry of Education and Science (ɍɎ03/03) and a grant from the University of Sofia (2005) have been used for calculations of mechanical strength of carbon nanotubes.
References Allan, M., 2005, Excitation of the four fundamental vibrations of CH4 by electron impact near threshold, J. Phys. B: At. Mol. Opt. Phys. 38:16791685. Baerends, E. J.; Ellis, D. E.; and Ros, P., 1973, Selfconsistent molecular Hartree—Fock—Slater calculations I. The computational procedure, Chem. Phys. 2:4151. Bienfait, M., Zeppenfeld, P., DupontPavlovsky, N., Muris, M.R., Wilson, T., DePies, M., and Vilches, O.E., 2004, Thermodynamics and structure of hydrogen, methane, argon, oxygen, and carbon dioxide adsorbed on sigle wall carbon naotube bundles, Phys. Rev. B 70:035410/110. Blöchl, P. E., 1994, Projector augmentedwave method, Phys. Rev. B 50:179531797. Blöchl, P. E., 2000, Firstprinciples calculations of defects in oxygendeficient silica exposed to hydrogen Phys. Rev. B 62:61586179. Car, R., Parrinello, M., 1985, Unified approach for Molecular Dynamics and DensityFunctional Theory, Phys. Rev. Lett. 55:24712474. Chelikowsky, J. R., 2000, The pseudopotentialdensity functional method applied to nanostructures, J. Phys. D: Appl. Phys. 33:R33R50 http://jrc.cems.umn.edu/codes.htm. Collins, P. G., Bradley, K., Ishigami, M., and Zettl, A., 2000, Extreme Oxygen Sensitivity of electron properties of carbon nanotubes, Science 287:18011804. Curtarolo, Stefano, 2003, CoarseGraining and Data Mining Approaches to the prediction of structures and their dynamics, Ph.D. Thesis, MIT (USA). Daykov, I., and Proykova, A., 2001, Cluster size dependence of the temperature interval of coexisting phases observed in plastic molecular clusters, in: Meetings in Physics at University of Sofia, A. Proykova, ed., vol. 2, Heron Press Science Series, Sofia, pp. 3135. Daykova, E., Pisov, S., and Proykova, A., 2005, Firstprinciples and molecular dynamics simulations of methane adsorption on graphene, NATO Science Series II, Springer, submitted. Daykova, E. , Proykova, A., and Ohmine, I., 2002, Statistical Calculations of Entropy of Methane Clathrate, in: Meetings in Physics @ University of Sofia, A. Proykova, ed., vol. 3, Heron Press Science Series, Sofia, pp. 44 48. Durgun, E., Dag, S., Bagci, V. M. K., Gulseren, O., Yildirim, T., and Ciraci, S., 2003, Systematic study of adsorption of single atoms on a carbon nanotube, Phys. Rev. B 67:201401201404.
206 Garao, C., Frontera, A., Quinonero, D., Costa, A., Ballester, P., Deya, P.M., 2003, Lithium diffusion in singlewalled carbon nanotubes: a theoretical study, Chem. Phys. Lett. 374:548553. Ghoniem, N., Busso, E., Kiossis, N., and Huang, H., 2003, Multiscale modeling of nanomechanics and micromechanics: an overview, Philosophical Magazine 83:34753528. Hertel, T., Kriebel, J., Moos, G., and Fasel, R., 2001, Adsorption and desorption of weakly bonded adsorbates from singlewall carbon nanotube bundles, AIP Conference Proceedings 590, Nanonetwork Materials, Fullerenes, Nanotubes and Related Systems, S. Saito, T. Ando, Y. Iwasa, K. Kikuchi, M. Kobayashi, Y. Saito, eds., Kamakura, Japan, p.181. Hill, T. L., 1987, Statistical Mechanics: Principles and Selected Applications, Dover, New York. Hill, T. L., 1986, An Introduction to Statistical Thermodynamic,.Dover, New York. Hill, T. L., 1994, Thermodynamics of Small Systems, Parts I and II, Two Volumes in One. Dover, New York. Iliev, H., and Proykova, A., 2004, Mechanically induced defects in singlewall carbon nanotubes, in Meetings in Physics at University of Sofia, A. Proykova, ed., vol. 5, Heron Press Science Series, Sofia, pp. 8791. Iliev, H., Li, F.Yin, and Proykova, A., 2005, Carbon nanotubes with vacancies under external mechanical stress and electric field, NATO Science Series II, Springer, submitted. Kadanoff, L. P., 1976, Notes on Migdal's Recursion Formulas, Ann. Phys. 100:359394. Kadanoff, L. P., 1977, The Application of Renormalization Group Techniques to Quarks and Strings, Rev. Mod. Phys. 49:267296. Kleinhammes, A., Mao, S.H., Yang, X.J., Tang, X.P, Shimoda, H., Lu, J.P., Zhou, O., and Wu, Y., 2003, Gas adsorption in singlewalled carbon nanotubes studied by NMR, Phys. Rev. B 68:075418/16. Kong, J., Franklin, N.R., Zhou, C., Chapiline, Peng, S., Cho, K., and Dai, H., 2000, Nanotube Molecular Wires as Chemical Sensors, Science 287:622625. Kohn, W. and Sham, L.J., 1965, Selfconsistent equations including exchange and correlation effects, Phys. Rev. 140:A1133A1138. Krasheninnikov, A.V., Nordlund, K., Lehtinen, P.O., Foster, A.S., Ayuela, A., and Nieminen, R.M., 2004, Adsorption and migration of carbon adatoms on carbon nanotubes: Densityfunctional ab initio and tightbinding studies, Phys. Rev. B 69:073402/14. Landman, Uzi and Kleinman, George, 1975, Local and nonlocal effects in the theory of physisorption, J. Vac. Sci. Technol. 12:206209. Lin, J. S. , Qteish, A., Payne, M. C. and Heine, V., 1993, Optimized and transferable nonlocal separable ab initio pseudopotentials, Phys. Rev. B 47:41744180. Maiti, A., Brabec, C.J., and Bernholc, J., 1997, Kinetics of metalcatalyzed growth of singlewalled carbon nanotubes, Phys.Rev. B 55:R6097R6100. Manfrin, R. (1996) On the gevrey regularity for weakly hyperbolic equations with spacetime degeneration of oleinik type, Rendiconti di Matematica 16:203231. Migdal, A.A., 1975, Zh.Eksp. Teor. Fiz. 69:1457. Moe, Zugang and Susan B. Sinnott, 2001, Separation of Organic Molecular Mixtures in Carbon Nanotubes and Bundles: Molecular Dynamics Simulations, The University of Kentucky, 177 Anderson Hall, Lexington, Kentucky 405060046. Muris, M., Bienfait, M., Zeppenfeld, P., DupontPavlovsky N., Wass, P., Johnsons, M., Vilches, O.E., and Wilson, T., 2001, http://www.ill.fr/AR01/p82.htm and references there in Nosé, S.J., 1984, Constant temperature molecular dynamics methods, J. Chem. Phys. 81:511519. Pati, R., Zhang, Y., and Nayak, S., 2002, Effect of H2O adsorption on electron transport in a carbon nanotube, Appl. Phys. Lett. 81:26382640.
207 Perdew, J. and Wang Y., 1992, Accurate and simple analytic representation of the electrongas correlation energy, Phys. Rev. B 45:1324413249. Perdew, J. P., Chevary, J. A., Vosko, S. H., Jackson, K. A. Pederson, M. R., Singh, D. J., and Fiolhais, C., 1992, Atoms, molecules, solids, and surfaces: Applications of generalized gradient approximation for exchange and correlation, Phys.Rev. B 46:66716687. Petrilli, H. M.; Blöchl, P. E.; Blaha, P.; Schwarz, K., 1998, Electricfieldgradient calculations using the projector augmented wave method, Phys. Rev. B 57:1469014697. Proykova, A., Pisov, S., Radev, R., Mihailov, P., Daykov, I., and R. S.Berry, 2001, Dynamical Phase Coexistence of Molecular Nanoclusters: Kinetics versus Thermodynamics, in: Nanoscience&Nanotechnolgy, E. Balabanova, I. Dragieva, eds., Heron Press Science Series, Sofia, pp.1820. Proykova, A., 2002, The role of metastable states in the temperature driven phase changes of molecular nanoclusters, Vacuum 68:97104. Proykova, A., Pisov, S., Dragieva, I., and Chelikowski, J., 2003, DFT Calculations of optimized structures of some coordination compounds, in: Nanoscience&Nanotechnolgy’03, E. Balabanova, I. Dragieva, eds., Heron Press Science Series, Sofia. Proykova, A., and Iliev, H., 2004, Simulated stress and stretch of SWCNT, in: Proceedings of SIMS, B. Elmegaard, J. Sporring, K. Erleben, K. Sorensen, eds., Copenhagen, pp. 273279. Shi, Wei and Johnson, J. Karl, 2003, Gas adsorption on Heterogeneous SingleWalled Carbon Nanotube Bundles, Phys. Rev. Lett. 91:015504/14. Skoulidas, A., Ackerman, D., Johnson, J.K., and Sholl, D., 2002, Rapid transport of gases in carbon nanotubes, Phys. Rev. Lett. 89:185901/14. Spurling, T. H., De Rocco, A. G., and Storvick, T. S., 1968, Intermolecular Forces in Globular Molecules. VI. Octopole Moments of Tetrahedral Molecules, J. Chem. Phys. J. Chem. Phys. 48:10061008. Talapatra, S. and Migone, A.D., 2002, Adsorption of methane on bundles of closedended singlewall carbon nanotubes, Phys. Rev. B 65:045416/16. Udomvech, A., Kerdcharoen, T., Osotchan, T., 2005, First principles study of Li and Li+ adsorbed on carbon nanotube: Variation of tubule diameter and length, Chem. Phys. Lett. 406:161166. Vassiliev, I., Ogut, S., and Chelikowsky, J. R., 2002, Firstprinciples densityfunctional calculations for optical spectra of clusters and nanocrystals, Phys. Rev. B 65:115416/118. Venables, John A., 2000, Introduction to Surface and Thin Film Processes, Cambridge University Press, Cambridge. Verlet, L., 1967, Computer "experiments" on classical fluids. I. Thermodynamical properties of LennardJones molecules, Phys. Rev. 159:98103. Zhang, Y. and Dai, H., 2000, Formation of metal nanowires on suspended singlewall carbon nanotubes, Appl. Phys. Let. 77:30153017. Zhang, Y., Franklin, N. W., Chen, R. J., and Dai, H., 2000a, Metal coating on suspended carbon nanotubes and its implication to metaltube interaction, Chem.Phys. Lett. 331:3541.
FIRSTPRINCIPLES AND MOLECULAR DYNAMICS SIMULATIONS OF METHANE ADSORPTION ON GRAPHENE EVGENIYA DAYKOVA, STOYAN PISOV, ANA PROYKOVA* Faculty of Physics, University of Sofia, Sofia, Bulgaria
Abstract. Firstprinciple realspace calculations within a timedependent local density approximation show fluctuating charge distribution in a methane molecule physisorbed on a finitesize graphene sheet, which in turn bends forming a concave surface under multipolemultipole interaction and the carbon network vibrates in a complex way: a metastable physisorbed state occurs.
Keywords: methane adsorption; graphene; Molecular Dynamics; abinitio calculation
This paper presents results from firstprinciples simulations of methane adsorption on graphene. One aim of the study is to quantify the difference in adsorption on a graphene sheet compared to adsorption on carbon nanotubes – perfect and defective. Comparisons show (Stan et al., 1998) a stronger binding of the adsorbate in the tubes than at the planar surface of graphite. Another goal is to find out if the adsorption makes a metastable state. Methane molecules (tetrahedra) resemble rare gases: CH4  CH4 interaction is classically described with a LennardJones potential, inherent for the noble atoms. Our previous Molecular Dynamics simulations showed that the interaction energy is orientationally dependent (Daykova et al., 2004). The current realspace density functional calculations with the Chelikowski’s code (Chelikowski, 2000) and an ultrasoft pseudopotential for carbon atoms generated with the fhi98PP code (Kratzer et al., 1998) do not manifest any dependence of the interaction energy on the CH4 orientation with respect to the initially flat graphene surface: case "F" with 3 bonds closer to the sheet; case "V" with 1 bond. The methane molecule is initially located in either "F" or "V" position above point 0 on the finite piece of a graphene sheet (Fig. 1a). After ~ 30 fs of relaxation, the CH4 center of mass stabilizes its position above the sheet *To whom correspondence should be addressed. Dr. Ana Proykova, Department of Atomic Physics, University of Sofia, 5 James Bourchier Blvd, 1164 Sofia, Bulgaria; email: [email protected]
209 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 209–210. © 2006 Springer. Printed in the Netherlands.
210 (Fig. 1b) (the upper curve is for case "F"). The interaction induces a dipole moment in CH4 (Fig. 1c) that fluctuates with the CH4 center of mass with a certain delay due to inertia. Larger moment is produced in case "V". The two orientations might be Raman distinguishable. The charge distribution at 20% level of the electron density 60 fs after the run start is presented in Fig. 1d. The preliminary results show that CH4 behaves differently on a tube.
(b)
(a)
(c)
(d)
Figure 1. Initial (a) and stabilized (b) position of the CH4 molecule above the graphene sheet, (c) induced dipole moment of the molecule, and (d) charge distribution of the relaxed configuration.
References Daykova, E., and Proykova, A., 2004, A peculiar structure of free methane clusters at low temperatures, Meetings in Physics @ University of Sofia, 5:6975. Chelikowsky, J. R., 2000, The pseudopotentialdensity functional method applied to nanostructures, J. Phys. D: Appl. Phys. 33: R33R50 http://jrc.cems.umn.edu/codes.htm. Kratzer, P., Morgan, C.G.,Penev, E., Rosa, A.L., Schindlmayr, A.,Wang, L.G.,Zywietz,T.,1998, Computer code for densityfunctional theory calculations for polyatomic systems, http://www.fhiberlin.mpg.de/th/fhi98md/fhi98PP/. Stan, G., and Cole, M. W., 1998, Low Coverage Adsorption in cylindrical pores, Surf. Sci. 395:280.
EFFECT OF SOLVENT AND DISPERSANT ON THE BUNDLE DISSOCIATION OF SINGLEWALLED CARBON NANOTUBES SILVIA GIORDANI,* SHANE D. BERGIN, ANNA DRURY, ÉIMHÍN NÍ MHUIRCHEARTAIGH, JONATHAN N. COLEMAN AND WERNER J. BLAU Molecular Electronics and Nanotechnology Group, Physics Department, Trinity College, Dublin 2, Ireland
Abstract. Singlewall carbon nanotubes (SWNT) are severely restricted in their applications, as they exist in ropelike bundles. Recently, J. Coleman et al. demonstrated a spectroscopic method to monitor bundle dissociation in low concentration NTpolymer composites.1 In this work we present the results we obtained with 2,5dioctyloxy1,4distyrylbenzene (pDSB), a shortchain analogues of the well known PPV. We found a strong dependence of the concentration at which individual SWNTs become stable with the nature of the solvent. Keywords: Singlewall carbon nanotubes; dispersion; debundling; fluorescence; microscopy
pDSB was synthesised in the house. It is strongly fluorescent, displays minimal selfaggregation and disperses SWNTs in the organic solvents. Stock solutions of pDSB and of SWCNT were made in the different solvents, and were subject to sonication. Two identical solutions of pDSB in each solvent were made and pure HiPCO SWNTs were added to one solution. Each solution was then diluted, serially, on the order of twenty times to give a broad range of CNT concentrations (102–108 kg/m3). Each solution was sonicated and then allowed to stand for 24 hours to come to equilibrium. No sedimentation was observed over this period. The absorption and emission spectra of each solution were recorded. The addition of the SWNTs to the pDSB results in a visible quenching of the emission, indicating interaction between the two. The molecule exists in two forms, free and bound to the SWNT. The PL efficiency for the bound form is expected to be extremely low as any photogenerated *To whom correspondence should be addressed. Silvia Giordani, Molecular Electronics and Nanotechnology Group, Physics Department, Trinity College, Dublin 2, Ireland; email: [email protected]
211 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 211–212. © 2006 Springer. Printed in the Netherlands.
212 singlet excitons preferentially decay nonradiatively through the fast vibrational manifold of the nanotubes. Thus any observed PL from the composite solutions is due to the free form only. For all concentrations, the fraction of free molecule in each solvent was calculated from the PL intensity ratio of composite solution to that of the pure molecule solution. Figure 1 reports the results obtained with pDSB in three different solvents. The solid black lines represent the response of an individualSWNT system as deduced from Coleman’s model, fitted for the results obtained in chloroform and in Nmethyl2pyrrolidone (NMP). In the case of the chlorinated solvent the experimental data deviates from this curve at SWNT concentrations higher than 1 x 107 kg/m3 indicating that bundles occur above this limit. In the case of the amide solvents the response does not deviate from the solid curve, indicating that individual SWNTs or small bundles of SWNTs are present up to the limit of the experiment. Preliminary AFM and TEM results agreed with the spectroscopic results. There is a clear shift in the bundle size distribution, to smaller bundles, in the amide solvent versus the chlorinated solvent. The dominant feature is the presence of small SWNT bundles at relatively high SWNT concentrations in the case of the amide solvent.
Figure 1. Graph of the fraction of free molecules as a function of concentration for pDSB in the different solvents.
References 1. J. N. Coleman, A. Fleming, S. Maier, S. O'Flaherty, A. Minett, M. S. Ferreira, S. Hutzler, W. J. Blau, Binding Kinetics and SWNT Bundle Dissociation in Low Concentration Polymer – CNT Dispersions, J. Phys. Chem. B 108, 34463450 (2004).
CARBON NANOTUBES WITH VACANCIES UNDER EXTERNAL MECHANICAL STRESS AND ELECTRIC FIELD HRISTO ILIEV, ANA PROYKOVA* Department of Atomic Physics, University of Sofia, Bulgaria FENGYIN LI National Chung Hsing University, Taiwan
Abstract.. Reduction of the gap in a (10,0) carbon nanotube with vacancies is obtained with the CASTEP 4.2 code. Molecular Dynamics simulations show that the vacancy distribution is important for the tube stability under stress.
Keywords: carbon nanotubes; defects; vacancies; electronic structure
The regular graphene structure that comprises carbon nanotubes (CNT) determines their extraordinary mechanical and electronic properties. In the presence of defects, the singlewall CNTs are less robust and their band structure can be changed. Understanding the role of the defects on the CNT properties is important for the development of CNT based nanodevices. The structural defects in CNTs can roughly be divided into topological defects and vacancies. We present a few results from Molecular Dynamics (MD) simulations of CNTs with different distributions and amounts of vacancies under an external load and a result from DFT calculations of the bandstructure of a singlewall (10, 0) CNT with ~ 1% vacancies. Classical MD simulations with a velocity Verlet algorithm for integration of the Newtonian equations of motion were performed to study defective CNTs response to an external field. Details of the simulations were published in the Ref. [1]. The simulations show that vacancies reduce the tensile strength of a SWCNT depending on their distribution and concentration as it is shown in Fig. 1. It might happen that different amounts of vacancies could have the same effect if they are properly distributed. *To whom correspondence should be addressed. Ana Proykova; email: [email protected]
213 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 213–214. © 2006 Springer. Printed in the Netherlands.
214 1.5
Strain energy [eV/atom]
1%
1.0 5%
0.5 10%
0.0 0.0
5.0
10.0
15.0
20.0
Time [ps]
Figure 1. The strain energy per atom in a (10,0) SWCNT with 2000 atoms and various defect concentrations (1%, 5% and 10%). For every concentration we present several curves corresponding to different defect distributions generated randomly.
We have also demonstrated with DFT calculations (CASTEP code) that a zigzag (10,0) SWNT with a monovacancy defect responds quite differently to a transverse electric field applied in different directions. The defect site of the defective nanotube provides a conducting band minimum with a lower energy (0.19 eV) than in the corresponding perfect nanotube (0.4 eV). Our results indicate that the band gap can be modulated by rotating a SWNT with a monovacancy defect under a transverse electric field. Upon application of a strong electric field, the band gap becomes similar to that found in the corresponding perfect nanotubes. This also provides some flexibility to select a suitable band gap for some specific applications because the band gap can increase or decrease depending on the direction and strength of the transverse electric field.
References 1.
Hristo Iliev and Ana Proykova, Mechanically induced defects in singlewall carbon nanotubes, in: Meetings in Physics @ University of Sofia, edited by A. Proykova, vol. 5 (Heron Press, Sofia, 2004) pp. 8791.
Part V. Mechanical properties of nanotubes and composite materials
MECHANICAL PROPERTIES OF THREETERMINAL NANOTUBE JUNCTION DETERMINED FROM COMPUTER SIMULATIONS ELENA BELOVA,* LEONID A. CHERNOZATONSKII Institute of Biochemical Physics of RAS, Kosygin St. 4, 119991 Moscow, Russia
Abstract. The mechanical properties of a threeterminal nanotube "bough" junction are described. Interatomic TersoffBrenner potential and intermolecular LennardJones potential were used to obtain the strain energy of the "bough" junction upon loading the "branch". At small strain, the energy increases quadratically with the strain and the "branch" reverts to its initial position without external force. For strains higher than the critical one, the junction falls to a metastable state “branch stuck to stem". Keywords: Nanotube junctions; Mechanical properties; Molecular dynamics simulations; TersoffBrenner potential
In this paper, the mechanical characteristics of a threeterminal (19,0)&(9,9) nanotube junction (Fig. 1a) are considered. This nanotube "bough" (B) junction1,2 consists of a "stem" nanotube with a "branch". The connection of two nanotubes is formed by defects (6 heptagons) introduced into the perfect hexagonal nanotube lattice. The topological defects create an angle of ~30° between the "stem" and the "branch". The atomic interactions were modeled by the TersoffBrenner potential3 and the "stem""branch" nanotube interaction was described by a LennardJones potential.4 The compression strain was created by a graphene plane parallel to the "stem" axis. The plane was shifted towards the "branch" by small steps and the junction was relaxed by a conjugategradient method keeping the "stem" ends constrained. The strain İ was defined as İ = (r0  r)/r0, where r0 is the "stem""branch" distance in initial state and r is the distance for "branch" under loading. The strain energy of this Bjunction for different strains was obtained. At strains İ < 0.5, the strain energy of junction increases as E(İ) = ½ E"İ2, where E" = 17 eV (Fig. 1c). In this case, the Bjunction is similar to a spring. *To whom correspondence should be addressed. Elena Belova, Institute of Biochemical Physics of RAS, Kosygin St. 4, 119991 Moscow, Russia; email: [email protected]
215 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 215–216. © 2006 Springer. Printed in the Netherlands.
216 The "branch" returns to its initial position after removal of the plane. If the strain is more than critical value İcr ~ 0.75, the Bjunction converts to a new metastable state "branch stuck to stem" (Fig. 1b). In this state, the "branch" remains parallel to the "stem" by van der Waals interactions. The distance between the "stem" and the "branch" is equal to the interlayer graphite distance at normal conditions of 0.34 nm. This new junction state is steady up to 2000 Ʉ heating with the "branch" oscillating near equilibrium position but not returning to the initial position.
Figure 1. A (19,0)&(9,9) Bjunction in (a) initial state and (b) "branch stuck to stem" state. (c) Strain energy vs. strain for the Bjunction.
In summary, our simulations show two possible types of behavior of threeterminal nanotube "bough"junction under compression strain. In the case of a small strain, this junction can be used as a "nanospring". However, we must take into account molecular effects appearing for atomicscale objects, e. g., the effect of "branch stuck to stem" in the nanotube Bjunction. This work is supported by the Russian Foundation for Fundamental Investigation (grants ʋ 050217443).
References 1. N. Gothard, C. Daraio, J. Gaillard, R. Zidan, S. Jin, and A. M. Rao, Controlled growth of Yjunction nanotubes using Tidoped vapor catalyst, Nano Lett. 4, 213217 (2004). 2. L. A. Chernozatonskii, Threeterminal junctions of carbon nanotubes: structures, properties and applications, J. Nanoparticle Res. 5, 473484 (2003). 3. D. W. Brenner, Empirical potential for hydrocarbons for use deposition of diamond films, Phys. Rev. B 42, 94589471 (1990). 4. S. B. Sinnott, O. A. Shenderova, C. T. White, and D. W. Brenner, Mechanical properties of nanotubule fibers and composites determined from theoretical calculations and simulations, Carbon 36, 19 (1998).
OSCILLATION OF THE CHARGED DOUBLEWALL CARBON NANOTUBE VICTOR LYKAH Dept. of Physics and Technology, National Technical University "Kharkiv Polytechnic Institute", Kharkiv 61002, Ukraine EVGEN S. SYRKIN* B. I. Verkin Institute for Low Temperature and Engineering, Nat'l Academy of Sciences of Ukraine, Kharkov 61103, Ukraine
Abstract. The oscillation frequency of charged doublewall semiconducting carbon nanotubes (DWCNTs) is found as function of the nanotube mass and geometry. The linear and nonlinear oscillations are analyzed.
Keywords: carbon nanotube; electronanomechanics; oscillation
Recent experiments in nanomechanic oscillations of the multiwall carbon nanotubes1,2 open perspectives in the Gigahertz range. The single molecule C60 transistor shows electromechanic oscillations.3 In this report, we consider the mechanical telescope oscillations of a DWCNT due to a charge (electron or hole) on it. The energy of the system is the sum of the Van der Waals interaction energy between the nanotubes (averaged and smoothened3) and the energy of the extra carrier interaction with the nanotube. For the nonresonant nanotube lengths, the ponderomotive electrostatic attractive forces give the main contribution to the potential of the oscillations of the DWCNT. The system geometry is illustrated in Fig. 1, when the length 2l of the external short tube is longer than the resonance one and a hole is localized on the external shorter tube. An electron is localized in the space between two tubes where the potential well is about 1 eV (Ref. [4]). The electrostatic energy of CNT is WNT 1 2 ³ H (r )H 0 dV , where H (r )H 0 is the tube electric permeability. For small shift [ of the short tube and for l , [ L , the nonlinear oscillation equation for the tube due to charge q is found as *To whom correspondence should be addressed: Evgen Syrkin; email: [email protected]
217 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 217–218. © 2006 Springer. Printed in the Netherlands.
218
1 1 1 q2 S (1 )[ ] 0 2 4 32S H 0 H ( L [ ) ( L [ )4 Here M is the oscillating tube mass, S is the longer tube crosssectional area. The expansion in series leads to the oscillation equation [ Z02[ [1 [ 2 / L2 ] 0 . M [
Figure 1 DWCNT geometry. The long (short) CNT has length 2L (2l). The distance between the middles of the tubes is ȟ. The arrows show the direction of the telescopic oscillation. The shorter CNT can be external (for holes and electrons) or internal (only for electrons).1,2
The frequency Z0 of DWCNT oscillation is a function of tube parameters Z # q 2 / 4S 2H 0 U (1 1/ H )(1/ L6 ) and Z02S # Z02L ( N L / N S ) . Here L(S) is for moving long (short) tube; M # 2S rLd U , S # 2S rd is used. U # m0 / da 2 is the mass density. m0 and a are carbon atom mass and the CC bond length. N L ~ rL L ( N S ~ rS l ) is number of atoms in the tube. The evaluation yields Z0 L ~ 108 y 1011 rad/s for H # 6 , L ~ 10 y 100 ǖ. The moving short tube can have considerably higher frequencies because N L !! N S . For relatively larger amplitudes, the nonlinearity gives rise to higher stiffness: Z 2 Z02 (1 [ 2 / L2 ) . The considered setup differs qualitatively from Refs. [1,2] where pure mechanical tube oscillations need large amplitude and extruding of tube core. 2 0L
References 1. S. B. Legoas, V. R. Coluci1, S. F. Braga, P. Z. Coura, S. O. Dantas, and D. S. Galvao, Molecular Dynamics Simulations of Carbon Nanotubes as Gigahertz Oscillators. Phys. Rev. Lett. 90, 055504, 2003. 2. Q. Zheng, J. Z. Liu, Q. Jiang. Excess van der Waals interaction energy of a multiwalled CNT with an extruded core and the induced core oscillation. Phys. Rev. B 65, 245409, 2002. 3. H. Park, J. Park, A. K. L. Lim, E. H. Anderson, A. P. Alivisatos, Paul L. McEuen. Nanomechanical oscillations in a singleC60 transistor. Nature 407, 57, 2000. 4. C. L. Kane, E. J. Mele, and A. T. Johnson. Theory of scanning tunneling spectroscopy of fullerene peapods. Phys. Rev. B 66, 235423, 2002.
POLYMER CHAINS BEHAVIOR IN NANOTUBES: A MONTE CARLO STUDY K. AVRAMOVA,* A. MILCHEV Institute for Physical Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria
Abstract. We study the equilibrium properties of flexible polymer chains confined in a soft tube by means of extensive Monte Carlo simulations.
Keywords: Monte Carlo computer simulations; polymer chains; elastic nanotubes; scaling; chain dynamics
The confinement of polymers in cylindrical tubes has attracted considerable attention by researchers during the last two decades in view of the broad class of applications such as protein transport through channels in membranes, etc. However, the behavior of linear polymers under conditions of geometric constraints is also a challenge for the basic research. We used an offlattice beadspring model for the MonteCarlo (MC) simulations with periodic boundary conditions in a box of size 128x128x128. The linear polymer chain consists of N = 32 to 256 effective monomers in a hexagonalcoordinated tube with 246 stacked rings in the box with Did = 3.08 to 10.9 in units of the maximum bond length (Fig. 1).
Figure 1. Snapshots of a polymer chain of length N = 128 in a tube of diameter Did = 3.08, 6.12, and 10.9 (from left to right) with hard (above) and soft (below) walls. *To whom correspondence should be addressed. Kati Avramova, Institute for Physical Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria; email: [email protected]
219 V.N. Popov and P. Lambin (eds.), Carbon Nanotubes, 219–220. © 2006 Springer. Printed in the Netherlands.
220
Figure 2. Profiles of the swollen region of the soft tube with radius R(Z) containing polymer chains of different length N = 32, 64, 128, and 256 along the tube axis Z. The four graphs correspond to four different diameters Did of unperturbed tubes containing no polymer.
Figure 2 shows a side view of the tube wall deformation in the vicinity of the enclosed polymer chain centered at the center of mass of the chain whereby the z axis is scaled proportionally to the number of monomers N in the polymer. Four tube diameters Did and data for chains with 32 < N < 256 are displayed. A good collapse of all curves on a cigarlike master profile of width W v N is observed. The sides of the bulge are almost perpendicular to the tube axis whereas its plateau is almost independent of N, at least for the narrow tubes, D0/Rg